CN-121998212-A - Logistics path collaborative optimization method and system

CN121998212ACN 121998212 ACN121998212 ACN 121998212ACN-121998212-A

Abstract

The invention discloses a logistics path collaborative optimization method and system, and relates to the technical field of logistics. The logistics path collaborative optimization method and system comprises the steps of S1, receiving user order tasks, collecting distribution data in real time, synchronizing distribution resource states, carrying out real-time evaluation of task and resource distribution, S2, establishing a traffic map network, marking distribution conflict points, generating distribution paths, carrying out unmanned vehicle and distributor path collaborative optimization, S3, continuously sensing distribution resource states, carrying out prediction analysis of availability and execution capacity of distribution resources, and implementing optimization measures, S4, identifying emergency events, carrying out task transfer and redistribution, dynamically readjusting distribution resources and paths, carrying out redistribution pressure evaluation, S5, analyzing system log records, constructing training data sets, evaluating scheduling effects and optimization strategies, and generating reports. The problem of delivery resource conflict caused by simultaneous operation of unmanned delivery vehicles and manual delivery operators in cities is solved.

Inventors

CHU ZILONG
SUN WEN
SUN XIAOYUAN
WU JIADONG
Sun Shenxiang
CHU ZIZHEN

Assignees

太川鲲(天津)物流科技有限公司

Dates

Publication Date: 20260508
Application Date: 20251225

Claims (10)

1. The logistics path collaborative optimization method is characterized by comprising the following steps of: S1, receiving a user order task, collecting distribution data in real time, synchronizing distribution resource states, and carrying out real-time evaluation of task and resource distribution; s2, a traffic map network is established, potential delivery conflict points are marked, a delivery path is preliminarily generated, and the unmanned vehicle and the dispatcher path are cooperatively optimized in real time according to the path conflict risk evaluation value; s3, continuously sensing the state of the distribution resources, carrying out predictive analysis on the availability and execution capacity of the distribution resources, and implementing optimization measures according to analysis results; S4, identifying an emergency, performing task transfer and reassignment, dynamically readjusting distribution resources and paths, and performing reassignment pressure evaluation; S5, analyzing system log records, constructing a training data set, evaluating the scheduling effect and the optimization strategy, and generating a report.
2. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of receiving a user order task, collecting distribution data in real time and synchronizing distribution resource status comprises the steps of: The method comprises the steps of receiving user order information through user application software and a merchant system interface, wherein the order information comprises a starting position, a terminating position, a delivery time limit, an article type comprising fragile articles, heavy articles and a cold chain, ordering time and customer preference, collecting delivery data in real time, wherein the delivery data comprise the steps of collecting unmanned vehicles and delivery person state data in real time through vehicle-mounted Internet of things terminal equipment and delivery person mobile phone software, the unmanned vehicles electric quantity, the delivery resource real-time position, the real-time delivery speed, the delivery load and the delivery resource availability, obtaining unmanned vehicle equipment energy consumption parameters through a vehicle-mounted sensor, collecting delivery person weight data, obtaining the current delivery person physical index through comprehensive analysis of a metabolic equivalent method based on delivery person weight data and the real-time delivery speed, synchronizing all unmanned vehicles and delivery person state data in real time, uploading all order information and delivery data to a delivery database, and analyzing and summarizing the total number of resources and the total number of tasks in real time.
3. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of performing real-time assessment of task and resource allocation is: Acquiring an order starting position and a real-time position of a distribution resource, and obtaining the distance between a current resource i and a task j starting position by using a map service interface; obtaining a total distribution distance according to a distance between a current resource i and a starting position of a task j and an order termination position, and calculating a total distribution time based on a real-time distribution speed of the distribution resource, obtaining energy consumption of an unmanned vehicle resource i for executing the task j through linear regression, obtaining energy consumption of the unmanned vehicle resource i for executing the task j through a metabolism equivalent method based on weight data of a distribution person and the total distribution time, calculating a total time required for completing the current task according to the starting position and the termination position of the order, obtaining a user preference in order information to obtain an expected delivery time of a user, subtracting the expected delivery time of the user from a sum of the total distribution time and the total time required for completing the current task, multiplying the distance between the current resource i and the starting position of the task j by a distance weight factor, multiplying the total distribution time by the weight factor, multiplying the energy consumption of the resource i for executing the task j by the energy consumption weight factor, summing up the four multiplication results to obtain a total item, traversing all resources and task allocation combinations, and if the resources i are assigned with a decision-making value of the task 1 and the total time is not assigned with a total time-consuming value of the task 1, and if the task execution time is not assigned with a total time-consuming value of the task is not assigned with a decision-making value of 0; And comparing the task resource allocation evaluation value with the allocation threshold value in real time, triggering a task re-allocation mechanism to automatically find a better combination if the task resource allocation evaluation value is greater than or equal to the allocation threshold value, implementing resource elimination and current limiting if a certain resource is continuously higher than the allocation threshold value, marking the resource elimination and current limiting as task failure, recording the task failure in a log, outputting an allocation result to an allocation decision table and storing the allocation result in an allocation buffer queue.
4. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of creating a traffic map network, marking potential distribution conflict points, and initially generating the distribution path is as follows: The method comprises the steps of establishing a basic road network structure based on an urban road network, accessing an urban traffic open platform and a camera sensor to acquire traffic data, carrying out real-time weighting to obtain road section congestion level and road section passing time, analyzing historical task tracks in real time based on historical order information, distribution data and historical distribution results, identifying high-frequency intersection areas of unmanned vehicles and distribution operators by using density clusters, marking areas with high overlapping probability of the unmanned vehicles and the distribution operators, training a graph neural network model by using the historical task tracks and the traffic data according to the basic road network structure, inputting distribution starting positions and ending positions to predict estimation of each node to a target ending position through the graph neural network model, considering the areas with high-frequency intersection and overlapping probability of the unmanned vehicles and the distribution operators in real time, guiding a path generation process to preferentially select adjacent nodes with high estimation, and finally outputting a node sequence of a primarily optimized task path.
5. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of collaborative optimization of the unmanned vehicle and the dispatcher path according to the path collision risk assessment value is as follows: Obtaining the node sequence of a task path output in the step of generating a delivery path, analyzing the node sequence of the task path to obtain the overlapping quantity of the path points of the task j and the task k, taking one path node number with the largest node number as the total number of the path points of the two tasks according to the node sequence of the task j and the task k, obtaining the real-time delivery speed of an unmanned vehicle and a delivery person, obtaining the speed difference of two delivery resources of the delivery task j and the task k at the same position according to the delivery speed difference of the node sequence at the overlapping point of the unmanned vehicle and the delivery person, obtaining the average delivery speed according to the real-time reporting of a vehicle-mounted Internet of things terminal device interface and the mobile phone software of the delivery person, obtaining the average delivery speed, obtaining the traffic density score of a road section, multiplying the overlapping quantity of the path points of the task j and the task k by the path overlap weight factor, multiplying the speed difference of the two delivery resources at the same position by the speed difference weight factor of the path overlap weight factor, obtaining the traffic density score and the three risk evaluation results; When the path conflict risk assessment value is smaller than the risk threshold value, the path conflict risk is considered to be lower, the distribution task is normally executed according to the current path, the path broadcasting is maintained, the current path is synchronized to the terminal equipment, and the normal scheduling flow is entered; When the path conflict risk assessment value is larger than or equal to the risk threshold value, a path buffer adjustment mechanism is triggered, conflict nodes and dense nodes in the path are identified, a buffer time window is set for each key node, sliding adjustment in a buffer area is allowed in the path execution process, when the path conflict is unavoidable, one minute delay and one-stop emergency measure are automatically introduced, a reprofiling module is called to generate an alternative route, a high-density and high-coincidence area is avoided, current limiting control and task transfer strategies are implemented on high-frequency conflict resources, and conflict point information is recorded into a high-risk node database.
6. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of performing predictive analysis of availability and execution capacity of distribution resources by continuously sensing the status of the distribution resources is as follows: Acquiring unmanned vehicles and dispatcher state data, constructing a state feature vector as a training set of a long-short-period memory network model, extracting, encoding and aligning key features of multi-type data of dispatching resources such as real-time positions, unmanned vehicle electric quantity, real-time dispatching speed and dispatching load, converting the key features into corresponding state feature vectors, integrating the corresponding state feature vectors into the state feature vectors, predicting the real-time state by using the long-short-period memory network model, predicting whether service capacity is reduced within a period of time, and outputting service availability scores from 0 to 1 according to the state prediction.
7. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of implementing the optimization measure according to the analysis result is as follows: Obtaining a service availability score of a long-short-period memory network model prediction output, obtaining a service availability score of a current moment, obtaining unmanned vehicle state data, obtaining current unmanned vehicle electric quantity, obtaining historical unmanned vehicle electric quantity and a distribution member physical ability index according to a distribution database, selecting a lowest numerical value, fitting by using a quantile regression algorithm to obtain the lowest acceptable electric quantity and the physical ability index of resources, obtaining a current distribution load, obtaining a distribution load according to the distribution database, obtaining the distribution load with the largest distribution success probability by using a logistic regression algorithm through fitting analysis, multiplying the service availability score of the current moment by a historical availability weight factor, subtracting the ratio of the current unmanned vehicle electric quantity and the distribution member physical ability index to the lowest acceptable electric quantity and the physical ability index of the resources by a constant, multiplying the energy influence weight factor, subtracting the ratio of the current distribution load to the maximum bearing capacity of the resources by the constant, multiplying the load influence weight factor, and summing the three product results to obtain a dynamic update value of the resource availability after a period of the future; The method comprises the steps of comparing a dynamic update value of resource availability with an availability threshold in real time, representing a low availability state when the dynamic update value of the resource availability is smaller than the availability threshold, automatically suspending resource order receiving qualification, triggering a task reallocation mechanism, transferring a task to a high availability resource, and simultaneously avoiding low availability resource from participating in a conflict intensive road section by a path cooperation and conflict avoidance module; And when the dynamic update value of the resource availability is greater than or equal to the availability threshold value, the dynamic update value of the resource availability is expressed as a high availability state, the task matching and path optimization are continuously participated, the light-load and short-distance tasks are preferentially matched, and the dynamic update value of the resource availability is recorded.
8. The method for collaborative optimization of physical distribution paths according to claim 1, wherein the specific process of identifying an emergency, performing task transfer and reassignment, dynamically readjusting distribution resources and paths, and performing reassignment pressure evaluation is as follows: The real-time monitoring abnormal indexes comprise that the electric quantity of an unmanned vehicle is lower than an electric quantity threshold value, resources continuously stand still at non-task nodes and exceed a parking threshold value, a risk score calculated based on the total distribution time and the user expected time in user preferences exceeds the risk threshold value, all available resources in the current city are reported and inquired in real time according to a vehicle-mounted Internet of things terminal equipment interface and the mobile phone software of a distributor, tasks of blocked resources are re-matched based on task resource distribution evaluation values, whether the tasks violate service standards are judged before re-matching, re-matching is canceled and marked as task failure is recorded in a log if the tasks violate the service standards, the real-time position of the current resources and the initial position of target tasks are acquired, a path cooperation and conflict avoidance module is called to regenerate a planning path, and all re-scheduling actions are recorded in the log; Obtaining and counting task rescheduling times recorded in a log to obtain the number of tasks to be rescheduled currently, obtaining high available resources with the resource availability dynamic update value being greater than or equal to the availability threshold value, counting to obtain the number of available resources currently, counting records marked as task failures in the scheduled log records to obtain the number of task failures in a current time period, counting the number of all order task allocation records in a period through an allocation decision table in a scheduling cache queue to obtain the total number of scheduled tasks in the current time period, adding the ratio of the number of task failures in the current time period to the total number of scheduled tasks in the current time period by a constant one, and multiplying the ratio of the number of tasks to be rescheduled currently to the number of available resources currently to obtain a task reallocation pressure value; Comparing the task reassignment pressure value with the primary pressure threshold and the secondary pressure threshold in real time, judging the normal state when the task reassignment pressure value is smaller than or equal to the primary pressure threshold, and executing the task reassignment according to a standard strategy; When the task reassignment pressure value is larger than the primary pressure threshold and smaller than or equal to the secondary pressure threshold, judging that the task reassigns the pressure value to be in a slight tension state, preferentially avoiding a path conflict area, prolonging a scheduling interval to relieve pressure, and controlling a new order access rate; When the task reassignment pressure value is larger than the secondary pressure threshold value, the task reassignment pressure value is judged to be in a high-pressure state, a standby resource pool, a task degradation mechanism and a regional current limiting strategy are immediately started, the high-frequency failure task is marked as a scheduling risk, and the high-frequency calling resource automatically enters a cooling state.
9. The method for collaborative optimization of a physical distribution path according to claim 1, wherein the specific process of analyzing a log record of a system, constructing a training data set, evaluating a scheduling effect and an optimization strategy, and generating a report is as follows: The method comprises the steps of continuously collecting key log information, constructing a high-quality training set based on collected delivery data, marking whether task allocation is successful, whether paths conflict and whether resources are high-efficiency multi-dimensional result labels through labeling processing, predicting task scheduling effects by adopting an integrated regression model, introducing a classification model to identify high-risk resources and task types, automatically generating a strategy execution effect report form each month, carrying out multi-dimensional performance monitoring and strategy duplication by combining task completion rate, failure rate and resource utilization rate indexes, and feeding back evaluation results to a terminal.
10. The logistics path collaborative optimization system is characterized by comprising a task receiving and distributing module, a path collaborative and conflict avoiding module, a resource state sensing and predicting module, a dynamic scheduling and rescheduling module and a data learning and feedback module: The task receiving and distributing module is used for receiving user order tasks, collecting distribution data in real time, synchronizing distribution resource states and carrying out real-time evaluation on task and resource distribution; The path coordination and conflict avoidance module is used for establishing a traffic map network, marking potential delivery conflict points, preliminarily generating a delivery path, and carrying out path coordination optimization of the unmanned vehicles and the delivery operators in real time according to the path conflict risk evaluation value; The resource state sensing and predicting module is used for continuously sensing the state of the distributed resources, performing predictive analysis on the availability and execution capacity of the distributed resources, and implementing optimization measures according to analysis results; the dynamic scheduling and rescheduling module is used for identifying an emergency, carrying out task transfer and reassignment, dynamically rescheduling distribution resources and paths and carrying out reallocation pressure evaluation; The data learning and feedback module is used for analyzing the system log records, constructing a training data set, evaluating the scheduling effect and the optimization strategy and generating a report.

Description

Logistics path collaborative optimization method and system Technical Field The invention relates to the technical field of logistics, in particular to a logistics path collaborative optimization method and a logistics path collaborative optimization system. Background Along with the continuous increase of urban logistics demands, the terminal distribution task has the characteristics of high density, diversification and strong timeliness, and particularly in complex scenes such as business circles, residential areas, office parks and the like, the distribution task has obvious space aggregation and time conflict. Meanwhile, the distribution resources are gradually expanded to multiple carriers such as intelligent unmanned vehicles, unmanned distribution robots and the like by traditional manual riders, and the scheduling problem is more complicated due to the isomerism of the resource states. In order to improve the overall logistics efficiency and reduce the distribution cost, an intelligent system capable of realizing multi-task and multi-resource collaborative matching and path optimization in a dynamic environment is constructed, and the intelligent system has become an important development direction of urban intelligent logistics. For example, the invention patent with the publication number of CN115187169A discloses a logistics distribution system and a logistics distribution method based on collaborative path planning, and relates to the technical field of supply chain logistics. The system comprises a data information acquisition module, a data preprocessing module, a delivery path planning module and a delivery instruction generation module, wherein warehouse information and customer order information are firstly acquired, the distance between each node is calculated, an omnidirectional network diagram is generated, customers are distributed to the warehouse according to the order information, a collaborative path planning model is solved to generate a delivery path and a transfer node, article delivery and transfer information is calculated, and finally a delivery instruction is sent to delivery equipment, so that the delivery equipment delivers articles in the warehouse to the customers. For example, the invention patent with publication number of CN114358675B discloses a multi-unmanned-plane-multi-truck collaborative logistics distribution path planning method, and belongs to the technical field of unmanned plane logistics. The method comprises the steps of firstly establishing a mixed integer linear programming model of a multi-unmanned aerial vehicle-multi-truck collaborative logistics distribution path programming problem, secondly carrying out initial programming on a truck distribution path based on a K-Means algorithm and a genetic algorithm, thirdly designing a path programming search operator, introducing a variable neighborhood search framework to jointly optimize the distribution paths of the unmanned aerial vehicle and the truck on the basis of the truck distribution path, and solving the established mixed integer linear programming model. According to the method provided by the invention, different purchase costs of the trucks and the unmanned aerial vehicles are considered, the purchase quantity and the use quantity of the transportation means are reasonably optimized, the total distribution cost is effectively reduced, the unmanned aerial vehicle-truck distribution scheme is further optimized, and the defects of the existing combined distribution model and method are overcome. However, in the process of implementing the technical scheme of the embodiment of the application, the application discovers that the above technology has at least the following technical problems: The matching mode is static, so that the resource allocation is difficult to adjust in real time, the path planning ignores conflict and traffic jam among resources, scheduling failure is easy to cause, the resource availability prediction is lacking, risk resources cannot be identified in advance, the data feedback mechanism is lacking, the strategy is difficult to optimize, the global optimization is difficult to realize under multi-objective coordination, and the intellectualization and practicability of the system are limited. Therefore, in view of the above problems, there is a need for a method and a system for collaborative optimization of a physical path. Disclosure of Invention Technical problem to be solved Aiming at the defects of the prior art, the invention provides a logistics path collaborative optimization method and a logistics path collaborative optimization system, which solve the problem of conflict of distribution resources caused by simultaneous operation of unmanned distribution vehicles and manual distribution operators in cities. Technical proposal The logistics path collaborative optimization method and system comprises the following technical scheme that S1, user order tasks are receive