Intelligent order-picking method for same-route multi-source logistics demand based on vehicle positioning

Abstract

This invention relates to the field of logistics transportation scheduling technology, specifically to an intelligent grouping method for multi-source logistics demands along the same route based on vehicle positioning. The method includes: receiving logistics delivery requests from multiple shippers, including delivery location, receiving location, cargo specifications, and desired time periods; acquiring real-time location, cargo space, and verified load capacity of available freight vehicles; identifying geographically overlapping request combinations through many-to-many route similarity calculation; aggregating cargo specifications and matching vehicle attributes to filter target vehicle sets; calculating the estimated travel time for vehicles to reach each delivery location; and generating a grouping transportation plan containing specific vehicles, pickup order, and time nodes based on the desired time period. This method optimizes the multi-source logistics grouping logic through route similarity analysis and real-time location time calculation, making the grouping plan more aligned with actual transportation and delivery needs, and improving the scheduling rationality of collaborative transportation of logistics demands along the same route.

Description

Technical Field

[0001] This invention relates to the field of logistics transportation scheduling technology, and in particular to a method for intelligent grouping of multi-source logistics demand along the same route based on vehicle positioning. Background Technology

[0002] Traditional logistics consolidation scheduling relies heavily on regional divisions of shipping and receiving locations to aggregate orders. It involves manually verifying or using basic information systems to match the total cargo volume with vehicle cargo space and rated load parameters. The scheduling process only focuses on the basic fit between vehicle empty status and cargo capacity, without detailed analysis of the route relationships between multiple logistics requests. The vehicle scheduling stage only matches cargo with vehicle capacity, failing to plan the receiving process based on real-time vehicle location. The consolidation plan merely establishes the correspondence between cargo and vehicles, lacking detailed planning for receiving time and sequence.

[0003] Existing group-buying models cannot accurately determine the geographical overlap between multiple logistics delivery requests. Order aggregation methods have limitations, easily leading to redundant transportation route planning and underutilization of vehicle resources. Vehicle scheduling does not incorporate real-time location calculations to determine arrival times at each delivery point, making it difficult to match group-buying results with the shippers' desired timeframes. The timing and execution sequence of the pickup process lack rationality, failing to meet the actual needs of collaborative transportation among multiple shippers on the same route. It is necessary to identify spatially overlapping request combinations through many-to-many route similarity calculations, while simultaneously calculating the estimated travel time at each delivery point based on real-time vehicle locations. This, combined with the desired timeframe, generates a complete group-buying transportation plan including vehicle information, pickup sequence, and time nodes. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning, comprising: Receive logistics shipment requests from multiple different shippers. The logistics shipment requests include the shipping location, receiving location, cargo specifications, and expected time period information. The system can obtain the real-time location and attribute information of available freight vehicles within the area. The vehicle attribute information includes cargo space and rated load capacity. Based on the shipping and receiving locations of all logistics delivery requests, many-to-many route similarity calculations are performed to identify request combinations with geographical overlap. The identified request combinations of cargo specifications are aggregated and matched with the obtained vehicle attribute information to filter out the target vehicle set that carries the aggregated cargo. Based on the real-time location of each vehicle in the target vehicle set, calculate the estimated travel time to each delivery location in each request combination; By combining the estimated travel time with the expected time period of each logistics delivery request, at least one feasible multi-source group-buying transportation plan is generated for each request. The multi-source group-buying transportation plan includes specific vehicles, picking order and time nodes.

[0006] As a further aspect of the present invention, the step of performing many-to-many route similarity calculation based on the shipping and receiving locations of all logistics delivery requests to identify request combinations with geographical overlap specifically includes: Map the shipping and receiving locations of all logistics shipment requests to a sequence of geographic coordinates, with each request forming a transportation line segment pointing from the shipping coordinates to the receiving coordinates; Calculate the straight-line distance between the starting points and the ending points of any two transport segments, as well as the average distance between the two segments in space; The calculated straight-line distance and average distance are input into the preset route overlap assessment model, which outputs a similarity score by combining the distance and the direction angle. Two logistics delivery requests with similarity scores higher than a set overlap threshold are grouped into the same set to be evaluated. A set to be evaluated may contain more than two requests. Based on the set to be evaluated, traverse all the shipping and receiving locations of all requests within the set to be evaluated, calculate the smallest convex polygon region covering all locations in the set to be evaluated, and mark the smallest convex polygon region as the collinear region of the request combination.

[0007] As a further aspect of the present invention, the step of aggregating the identified request combinations of cargo specifications and matching them with the obtained vehicle attribute information specifically includes: Extract the cargo specifications of all logistics shipping requests in a request combination. The cargo specifications include volume, weight and cargo type. The volumes of all goods in the requested combination are summed to obtain the total aggregate volume, and the weights of all goods are summed to obtain the total aggregate weight. Based on the type of goods, determine whether there are loading constraints, including the prohibition of mixing special goods, and perform virtual loading simulation on aggregated goods based on the judgment result; Iterate through the real-time acquired vehicle attribute information to filter out candidate vehicles whose cargo space is greater than the aggregated total volume and whose rated load is greater than the aggregated total weight. For the selected candidate vehicles, a second verification is performed based on their cargo box structure parameters and the results of virtual loading simulation. Vehicles that cannot meet the cargo type loading constraints are eliminated, and the remaining vehicles constitute the target vehicle set.

[0008] As a further aspect of the present invention, the step of calculating the estimated travel time to each delivery location in each request combination based on the real-time location of each vehicle in the target vehicle set specifically includes: Obtain the real-time latitude and longitude coordinates of each vehicle in the target vehicle set, and request the geographic coordinates of all delivery locations in the combined collinear area; Based on real-time traffic data, the optimal driving route from the real-time location of each vehicle to each delivery location is planned, taking into account the current road congestion. Based on the length of the planned optimal driving route and the real-time road traffic speed, calculate the travel time for each vehicle to reach each delivery location; Based on the travel time en route, a fixed vehicle preparation time and a loading operation baseline time are added to obtain the total estimated travel time from the vehicle's current location to the point where loading is completed. Construct an estimated travel time matrix, where the rows of the matrix correspond to the vehicles in the target vehicle set, the columns correspond to the delivery locations in the request combination, and the matrix elements are the corresponding total estimated travel time.

[0009] As a further aspect of the present invention, the step of combining the estimated travel time with the expected time period information of each logistics delivery request to generate at least one feasible multi-source group-buying transportation plan for each request specifically includes: For a request combination and its corresponding estimated time matrix, each row of the matrix, i.e. each vehicle, is used as the vehicle to execute the group-buying transportation plan. Starting with the vehicle to be executed, based on the time consumption data in the estimated time consumption matrix, a path search algorithm is used to find a sequence that visits all delivery locations of the requested combination and finally completes all deliveries. The sequence must satisfy that each location is visited only once. While searching the access sequence, calculate the actual arrival time of each shipping location in the sequence. The actual arrival time is equal to the departure time of the previous location in the sequence plus the travel time from the previous location to the current location. The actual arrival time of each shipping location is calculated and compared with the expected time period information of the logistics shipping request corresponding to the shipping location to check whether it falls within the expected time period. If there exists an access sequence such that the actual arrival time of all shipping locations meets their respective expected time periods, then the executing vehicle is bound to the access sequence to generate a complete group-buying transportation plan, in which the estimated operation time of each node is specified.

[0010] As a further aspect of the present invention, the step of employing a path search algorithm to find a sequence that visits all shipping locations in the request combination and ultimately completes all deliveries specifically includes: Set the search starting point as the real-time location of the vehicle being executed, and treat all shipping and receiving locations in the request combination as the set of nodes that must be visited. In the node set, the real-time location of the executing vehicle is introduced as a virtual starting point, and the last receiving location is introduced as a virtual ending point to construct a complete access graph model. In the access graph model, the weight of an edge is determined by the travel time between two points under real-time traffic conditions, and the service time of a node is determined by the loading and unloading time of goods. Using an improved saving or insertion algorithm, under the premise of satisfying vehicle load and volume constraints, a Hamiltonian path is constructed in the access graph model that starts from a virtual starting point, visits all nodes, and finally reaches a virtual ending point. The constructed Hamiltonian path is locally optimized by swapping the visit order of adjacent nodes or adjusting sub-paths to reduce the total travel time, and finally the optimized visit sequence is output.

[0011] As a further aspect of the present invention, the calculation of the minimum convex polygon region covering all location points of the set to be evaluated specifically includes: Read the geographic coordinates of the shipping and receiving locations of all logistics shipping requests in a set to be evaluated, and form a set of geographic coordinate points to be covered. Find the point with the smallest latitude from the set of geographic coordinate points. If there are multiple points, select the point with the smallest longitude as the starting point. Using the starting base point as the origin of polar coordinates, calculate the polar angles of all other points in the point set relative to the origin of polar coordinates, and sort them in order of increasing polar angles; Process each sorted point in turn, compare the current point with the last two points that have formed the boundary of the convex polygon, and determine whether it forms a convex edge or needs to backtrack to delete the concave point. After traversing all points, connect the finally determined points to form a convex polygon, which is the smallest convex polygon region covering all the points of the requested combination, and record the vertex coordinates of the smallest convex polygon.

[0012] As a further aspect of the present invention, the step of planning the optimal driving path from the real-time location of each vehicle to each delivery location based on real-time traffic data specifically includes: Connect to the city's real-time traffic information service platform to obtain current road network topology data, average driving speed of each road segment, and congestion event information; The road network is modeled as a weighted directed graph, where nodes are road intersections and the weights of edges represent the estimated travel time of a road segment under the current road conditions. In a weighted directed graph, the real-time location of a vehicle is mapped to the nearest road node as the starting point of the path, and the delivery location is mapped to the nearest road node as the ending point of the path. Run the shortest path algorithm on a weighted directed graph to find the sequence of nodes with the minimum cumulative weight from the starting node to the ending node, where the weight represents the shortest travel time. The node sequence found by the shortest path algorithm is mapped back to the actual road to generate a specific optimal driving path that takes into account real-time traffic conditions, and its total length and estimated travel time are recorded.

[0013] As a further aspect of the present invention, the calculation of the actual arrival time of each shipping location in the sequence specifically includes: Determine the start time of the group-buying transportation plan, where the start time is the initial moment when the vehicle departs from its real-time location; Read the generated access sequence. The first element of the sequence is the first delivery location to be accessed. Obtain the total estimated travel time from the real-time vehicle location to the first delivery location from the estimated travel time matrix. Add the initial time to the total estimated travel time to obtain the actual arrival time of the vehicle at the first delivery location; After obtaining the actual arrival time at the first shipping location, add the standard cargo loading operation time at that shipping location to obtain the time when the vehicle leaves the shipping location; Based on the time the vehicle leaves the previous position, and adding the travel time from the previous position to the next position in the sequence obtained from the estimated time matrix, the actual arrival time of the next position in the sequence is iteratively calculated until all positions in the sequence have been calculated.

[0014] As a further aspect of the present invention, the local optimization of the constructed Hamiltonian path, which reduces the total travel time by exchanging the access order of adjacent nodes or adjusting sub-paths, specifically includes: Obtain the Hamiltonian path initially constructed by the path search algorithm, which includes the access order of all nodes from the virtual starting point to the virtual ending point; A two-point swapping strategy is adopted to try swapping the access order of any two internal nodes that are neither the starting point nor the ending point in the path in turn, forming a new access sequence; For each newly generated access sequence, recalculate its total travel time under the current real-time traffic conditions, including the travel time between nodes and the service time on the nodes; If the total travel time of the new access sequence is less than the total travel time of the original Hamiltonian path, and does not violate the expected time period constraints of each logistics delivery request, then the new sequence replaces the original path; After completing all two-point swap attempts, a path segment reversal strategy is adopted. A continuous node segment in the path is randomly selected, and the access order of the path segment is reversed. The total travel time is evaluated to see if it is reduced. If it is reduced, the change is accepted. This process is repeated until the total travel time cannot be further optimized.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: The system calculates the route similarity between the shipping and receiving locations of many-to-many logistics requests, directly identifying request combinations with geographical overlap. This changes the traditional method of aggregating orders by single region, allowing the aggregation of multi-source logistics demands to be completed based on the spatial correlation of actual transportation routes. This reduces redundant parts of transportation route planning, ensuring that cargo combination methods match actual transportation routes and optimizing the route arrangement for collaborative transportation of multiple orders. Based on the real-time vehicle location, the system calculates the estimated travel time for vehicles to reach each shipping location within the request combination, transforming the actual vehicle location status into quantifiable travel data. This allows the consolidation process to fully consider the real-time distribution of vehicles, avoiding scheduling deviations caused by matching solely based on vehicle empty load attributes, and making the adaptation of vehicles and request combinations more closely aligned with actual transportation scenarios.

[0016] By combining estimated travel time with the expected timeframe of logistics delivery requests, multi-source group-buying transportation plans with specific pickup sequences and time nodes can be generated. This refines the execution elements of the group-buying plan, ensuring that transportation scheduling matches the shipper's time requirements. Clearly defining the specific vehicle information corresponding to the group-buying plan allows for more targeted execution of multi-source logistics needs, making the pickup and transportation process more aligned with actual transportation execution conditions. It also makes the scheduling of multi-source logistics needs along the same route more consistent with the operational logic of actual transportation scenarios, optimizing the scheduling and execution of multi-order collaborative transportation. Attached Figure Description

[0017] Figure 1 This is a flowchart of the intelligent group-buying method for multi-source logistics demand on the same route based on vehicle positioning, as described in this invention. Figure 2 A flowchart for aggregating cargo specifications and matching vehicle attributes. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1 The system receives logistics delivery requests from multiple different shippers. Each request includes the shipping location, receiving location, cargo specifications, and expected time period. It acquires real-time locations and vehicle attribute information of available freight vehicles within the region, including cargo space and rated load capacity. Based on the shipping and receiving locations of all logistics delivery requests, it performs many-to-many route similarity calculations to identify request combinations with geographical overlap. It aggregates the cargo specifications of the identified request combinations and matches them with the acquired vehicle attribute information to filter out a set of target vehicles capable of carrying the aggregated cargo. Based on the real-time location of each vehicle in the target vehicle set, it calculates the estimated travel time to each shipping location in each request combination. Combining the estimated travel time with the expected time period information of each logistics delivery request, it generates at least one feasible multi-source consolidation transportation plan for each request combination. This multi-source consolidation transportation plan includes specific vehicles, pickup order, and time nodes.

[0021] In one embodiment of the present invention, three received logistics delivery requests have their delivery and receipt locations geocoded and mapped to geographic coordinate sequences. The first logistics delivery request forms a transportation line segment from coordinate point A1 to coordinate point B1, the second logistics delivery request forms a transportation line segment from coordinate point A2 to coordinate point B2, and the third logistics delivery request forms a transportation line segment from coordinate point A3 to coordinate point B3. The straight-line distance between the starting points and ending points of the transportation line segments of the first and second logistics delivery requests, as well as the average distance between the two line segments in space, are calculated. These calculated straight-line distances and average distances are input into a preset route overlap assessment model, which outputs a similarity score by combining the distance and the angle between the directions of the two line segments.

[0022] The similarity score output by the route overlap assessment model is calculated using the following formula: in: Indicates the first The logistics shipment request and the first Route similarity score between individual logistics shipping requests; This represents the straight-line distance between the starting points of two transport segments; This represents the straight-line distance between the endpoints of two transport segments. This represents the average distance between two transport segments in space. This represents the angle between the direction vectors of two transport line segments; , , , These are preset weighting coefficients used to adjust the influence of different distance and direction factors on the similarity score.

[0023] In practice, a fixed overlap threshold is set, for example, 0.75. The calculated similarity score is compared with the set overlap threshold. If the similarity score between the first and second logistics shipping requests is higher than the overlap threshold, these two logistics shipping requests are grouped into the same evaluation set. Similarly, if the similarity score between the second and third logistics shipping requests is also higher than the overlap threshold, the third logistics shipping request is also included in this evaluation set, making the evaluation set contain three logistics shipping requests. An evaluation set can contain more than two logistics shipping requests.

[0024] Based on the set to be evaluated, traverse the shipping and receiving locations of all logistics shipping requests within the set, and calculate the minimum convex polygon region covering all locations in the set. Read the geographic coordinates of the shipping and receiving locations of all logistics shipping requests in this set; these coordinates form a set of geographic coordinate points to be covered, containing points A1, B1, A2, B2, A3, and B3. Find the point with the smallest latitude in the geographic coordinate point set; if there are multiple points with the smallest latitude, select the point with the smallest longitude and use this point as the starting base point. Using the starting base point as the origin of polar coordinates, calculate the polar angle of all other points in the point set relative to the origin, and sort all points in ascending order of polar angle. Process each sorted point sequentially, comparing the current point with the two points that already form the boundary of the convex polygon to determine whether the current point forms a convex edge or needs backtracking to delete concave points. After traversing all sorted points, connect the finally determined points; the resulting convex polygon is the minimum convex polygon region covering all locations in the set to be evaluated. Mark the smallest convex polygon region as the collinear region to be combined, and record the vertex coordinates of the smallest convex polygon.

[0025] In some embodiments, the weighting coefficients of the route overlap assessment model can be configured according to the road network characteristics of different urban areas. In urban centers, the weight of the directional angle factor can be set higher. In some embodiments, the average distance between two transport segments in space is calculated by uniformly sampling the two segments and calculating the average distance between all sampled point pairs. Optionally, the overlap threshold value can be dynamically adjusted according to the density of logistics orders and the real-time scheduling requirements. Optionally, when calculating the polar angle, for points with the same polar angle, only the point farthest from the origin of polar coordinates is retained to participate in the subsequent construction of the convex polygon. It can be understood that the calculation of the minimum convex polygon region provides a spatial aggregation range for subsequent vehicle route planning. Through the calculation of similarity scores and the minimum convex polygon region, suitable combinations of logistics delivery requests for merged transportation can be effectively identified from a geospatial perspective.

[0026] In one embodiment of the present invention, see [reference] Figure 2 An identified request combination contains three logistics shipping requests. The specifications of the goods in the first logistics shipping request are extracted, recording a volume of 0.8 cubic meters, a weight of 120 kg, and a goods type of "general daily necessities." The specifications of the goods in the second logistics shipping request are extracted, recording a volume of 1.2 cubic meters, a weight of 200 kg, and a goods type of "general daily necessities." The specifications of the goods in the third logistics shipping request are extracted, recording a volume of 0.5 cubic meters, a weight of 80 kg, and a goods type of "fresh and frozen goods." The volumes of all goods in this request combination are added together: 0.8 cubic meters + 1.2 cubic meters + 0.5 cubic meters = a total aggregate volume of 2.5 cubic meters. The weights of all goods in this request combination are added together: 120 kg + 200 kg + 80 kg = a total aggregate weight of 400 kg.

[0027] Based on the cargo type, the system determines whether loading constraints exist. In practice, a cargo type compatibility rule table is pre-set. This table defines that "fresh and frozen goods" cannot be mixed with "chemicals," but can coexist with "general daily necessities" under physical separation conditions. Based on this determination, virtual loading simulation is performed on the aggregated cargo. The simulation considers cargo shape and loading order; for example, "fresh and frozen goods" are placed at the rear of the cargo container near the hatch. Virtual loading compatibility is evaluated using the following formula: in: This indicates the virtual loading adaptation score; This represents the actual volume of cargo space that is planned to be occupied in the virtual loading simulation; This indicates the total volume of the vehicle's cargo space; This represents the total weight of the cargo that is actually planned to be loaded in the virtual loading simulation; Indicates the vehicle's approved load capacity; It is a coefficient based on cargo type compatibility rules. It takes a value of 1 when the mixed loading constraint is fully satisfied, a value of 0.8 when the constraint is partially satisfied, and a value of 0 when the constraint is violated.

[0028] The system iterates through real-time vehicle attribute information obtained from the vehicle-to-everything (V2X) platform, including cargo space and rated load capacity. Candidate vehicles with a cargo space greater than 2.5 cubic meters of aggregated total volume and a rated load capacity greater than 400 kg of aggregated total weight are selected. For example, Vehicle A has a cargo space of 3 cubic meters and a rated load capacity of 500 kg, and Vehicle B has a cargo space of 2.8 cubic meters and a rated load capacity of 450 kg; these two vehicles are included in the candidate vehicle list. Vehicle C has a cargo space of 2 cubic meters, and although its rated load capacity of 600 kg meets the requirements, it is not selected as a candidate vehicle because its cargo space is less than 2.5 cubic meters of aggregated total volume.

[0029] For the shortlisted candidate vehicles, a secondary verification was performed based on the cargo box structure parameters and the results of the virtual loading simulation. Vehicle A's cargo box structure parameters show that its cargo box is a stake type, lacking physical isolation. Although its volume and load capacity meet the requirements, in the virtual loading simulation, because it cannot provide independent temperature zone isolation for "fresh cold chain products," the calculated virtual loading suitability evaluation score is low. The score is 0, therefore vehicle A is eliminated. Vehicle B's cargo box structure parameters indicate it is equipped with a movable partition. In the virtual loading simulation, by planning the partition's position, independent space can be created for "fresh cold chain products," resulting in a virtual loading adaptability evaluation score. The value is 0.8, which satisfies the loading constraint, therefore vehicle B is retained. After a second verification, the remaining vehicles constitute the target vehicle set, which in this scenario includes vehicle B.

[0030] In some embodiments, cargo specifications may also include the length, width, and height dimensions of the cargo. The virtual loading simulation performs three-dimensional stacking optimization calculations based on these dimensions to determine whether the cargo can be loaded into the container. In some embodiments, cargo types also include "fragile goods," "precision instruments," etc., whose loading constraints may require securing, shockproofing, or prohibition of heavy pressure. Optionally, a leniency factor can be introduced into the calculation of the aggregate total volume, for example, multiplying the calculated aggregate total volume by 1.1 to reserve some loading margin. Optionally, when the cargo type includes "dangerous goods," the virtual loading simulation will forcibly check whether it is compatible with any other cargo type; if incompatible, the requested combination is directly determined not to be loaded together.

[0031] In one embodiment of the present invention, the real-time location latitude and longitude coordinates of each vehicle in the target vehicle set are obtained. The target vehicle set includes vehicle A and vehicle B. The real-time location latitude and longitude coordinates of vehicle A are (116.3000, 39.9000), and the real-time location latitude and longitude coordinates of vehicle B are (116.3200, 39.9200). The geographical coordinates of all delivery locations in the collinear area of ​​the request combination are obtained. The collinear area includes two delivery locations: the geographical coordinates of delivery location X are (116.3500, 39.9500), and the geographical coordinates of delivery location Y are (116.3400, 39.9300). Based on real-time traffic data, the optimal driving path from the real-time location of each vehicle to each delivery location is planned. The system connects to the city's real-time traffic information service platform to obtain the current road network topology data, the average driving speed of each road segment, and congestion event information. The road network is modeled as a weighted directed graph, where nodes represent road intersections and edge weights represent the estimated travel time of a route under current traffic conditions. In this graph, the real-time location of vehicle A is mapped to the nearest road node as the path's starting point, and the delivery location X is mapped to the nearest road node as the path's ending point. A shortest path algorithm is run on the graph to find the sequence of nodes with the minimum cumulative weight from the starting point to the ending point; the weight represents the shortest travel time. Data from the city's real-time traffic information service platform shows that on the current path from vehicle A to delivery location X, route AB is congested with an average speed of 15 km / h, while the alternative route CD is unobstructed with an average speed of 40 km / h. The shortest path algorithm will select the node sequence with the shorter cumulative travel time, i.e., the path that includes the alternative route CD. The node sequence found by the shortest path algorithm is mapped back to the actual road to generate a specific optimal driving route that takes into account real-time traffic conditions, such as "XX Road -> CD Elevated Road -> YY Road", and its total length of 8.5 kilometers and estimated travel time of 15 minutes are recorded.

[0032] In practice, the optimal route from vehicle A's real-time location to delivery location Y is planned, resulting in a 6-kilometer route with an estimated travel time of 10 minutes, calculated based on real-time traffic conditions. The optimal route from vehicle B's real-time location to delivery location X is also planned, generating a 5-kilometer route with an estimated travel time of 9 minutes. The optimal route from vehicle B's real-time location to delivery location Y is further planned, generating a 3-kilometer route with an estimated travel time of 6 minutes. Based on the length of the planned optimal routes and the real-time road speed, the transit time for each vehicle to reach each delivery location is calculated. In addition to the transit time, fixed vehicle preparation time and loading operation baseline time are added, with vehicle preparation time uniformly set at 10 minutes and loading operation baseline time uniformly set at 15 minutes. Calculate the total estimated travel time from the vehicle's current location to the loading point at the destination. The estimated travel time from vehicle A to destination X is 15 minutes + 10 minutes + 15 minutes = 40 minutes. The estimated travel time from vehicle A to destination Y is 10 minutes + 10 minutes + 15 minutes = 35 minutes. The estimated travel time from vehicle B to destination X is 9 minutes + 10 minutes + 15 minutes = 34 minutes. The estimated travel time from vehicle B to destination Y is 6 minutes + 10 minutes + 15 minutes = 31 minutes. (Travel time en route) The calculation formula is: in: This represents the travel time of a vehicle along path P from the starting point to the end point. This represents the optimal driving route planned by the shortest path algorithm, which consists of a series of continuous road segments; This represents a specific segment of the path P; Indicates the length of road segment e; This represents the real-time average speed of road segment e at calculation time t, and this data is obtained from real-time traffic conditions.

[0033] Construct an estimated travel time matrix. The rows of the matrix correspond to vehicles in the target vehicle set, and the columns correspond to delivery locations in the request combination. Each element of the estimated travel time matrix represents the corresponding total estimated travel time. Based on the calculation results, the resulting estimated travel time matrix is ​​as follows: rows represent vehicles A and B, and columns represent delivery locations X and Y. The element corresponding to delivery location X for vehicle A is 40, and the element corresponding to delivery location Y for vehicle A is 35. The element corresponding to delivery location X for vehicle B is 34, and the element corresponding to delivery location Y for vehicle B is 31.

[0034] In some embodiments, vehicle preparation time can be differentiated based on different vehicle models or conditions; for example, refrigerated trucks require additional time to start their refrigeration equipment. In some embodiments, the baseline loading operation time can be dynamically adjusted based on the historical average operational efficiency of the delivery location or the type of cargo; the baseline loading operation time for bulk cargo is longer than that for standard cargo. Optionally, real-time road traffic speed... The value can be the average speed of the road segment over the past 5 minutes, or a value that incorporates short-term predicted speeds for the future. Optionally, the travel time can be included before calculating the total estimated trip time. The system multiplies the estimated time by a dynamic reliability coefficient based on whether the current time falls within peak traffic hours, increasing the redundancy of the time forecast. This means that planning routes and calculating times based on real-time traffic data makes the estimated travel time more closely reflect actual road conditions.

[0035] In one embodiment of the present invention, referring to Table 1, a request combination includes two logistics delivery requests, corresponding to delivery location X and delivery location Y, and the corresponding estimated time matrix is ​​shown in Table 1, where the time unit is minutes.

[0036] Table 1: Estimated Travel Time Matrix (Vehicle to Delivery Location) In practical implementation, for this request combination and its corresponding estimated time matrix, each row of the estimated time matrix, i.e., each vehicle, is used as the execution vehicle for the group-buying transportation plan. First, vehicle A is selected as the execution vehicle for the group-buying transportation plan. Starting from the execution vehicle, based on the time data in the estimated time matrix, a path search algorithm is used to find a sequence that accesses all delivery locations and ultimately completes all deliveries. The access sequence must satisfy the condition that each location is visited only once. A possible access sequence output by the path search algorithm is "delivery location X -> delivery location Y". While searching for access sequences, the actual arrival time of each delivery location in the sequence is calculated. The start execution time of the group-buying transportation plan is determined, which is the initial time when the execution vehicle departs from its real-time location, set to 9:00 AM. The generated access sequence is read; the first element of the sequence is the first delivery location X to be visited. The total estimated travel time from the real-time location of vehicle A to the first delivery location X is obtained from the estimated time matrix: 40 minutes. Adding the initial time of 9:00 AM to the total estimated travel time of 40 minutes, the actual arrival time of vehicle A at the first shipping location X is calculated to be 9:40 AM. After obtaining the actual arrival time at the first shipping location X, adding the standard cargo loading operation time at shipping location X (set to 20 minutes), the departure time of vehicle A from shipping location X is calculated to be 10:00 AM. Based on the departure time of vehicle A from the previous location, adding the travel time from the previous location to the next location in the sequence obtained from the estimated travel time matrix (the travel time from shipping location X to shipping location Y is obtained by querying the path database and is 25 minutes), the actual arrival time of the next location in the sequence, shipping location Y, is calculated iteratively to be 10:25 AM. Calculating the departure time of shipping location Y, the actual arrival time of shipping location Y is 10:25 AM, plus the standard cargo loading operation time of 20 minutes at shipping location Y, yields the departure time of vehicle A from shipping location Y to be 10:45 AM. The formula for calculating the departure time of a shipping location is: in: Indicates that the vehicle has left the first Time at each location; Indicates that the vehicle has arrived at the The actual arrival time at each location; Indicates the vehicle is in The standard service time for loading or unloading cargo at each location.

[0037] The calculated actual arrival time for each shipping location is compared with the expected time slot information of the corresponding logistics shipping request. The expected time slot for the logistics shipping request for shipping location X is 9:30 AM to 10:30 AM, and the expected time slot for the logistics shipping request for shipping location Y is 10:00 AM to 11:00 AM. The actual arrival time of shipping location X at 9:40 AM falls within the expected time slot of 9:30 AM to 10:30 AM, and the actual arrival time of shipping location Y at 10:25 AM falls within the expected time slot of 10:00 AM to 11:00 AM. If there exists an access sequence that ensures the actual arrival times of all shipping locations meet their respective expected time slots, then the vehicle will be bound to the access sequence to generate a complete groupage transportation plan. The plan specifies the estimated operation time for each node. In this implementation, the access sequence "shipping location X -> shipping location Y" satisfies all time period requirements, thus generating a group-buying transportation plan with vehicle A as the execution vehicle. The specific time nodes included in the plan are as follows: vehicle A departs from the current location at 9:00 AM, is expected to arrive at shipping location X at 9:40 AM and begin loading, is expected to leave shipping location X at 10:00 AM, is expected to arrive at shipping location Y at 10:25 AM and begin loading, and is expected to leave shipping location Y at 10:45 AM to head to the receiving location.

[0038] In one embodiment of the present invention, the search starting point is set as the real-time location of the executing vehicle R. All delivery and receiving locations in the request combination are considered as a set of nodes that must be visited. The node set includes delivery location A, delivery location B, receiving location C, and receiving location D. Within this node set, the real-time location of the executing vehicle R is introduced as a virtual starting point S, and the last receiving location is introduced as a virtual ending point E, constructing a complete access graph model. In the access graph model, the edge weights are determined by the travel time between two points under real-time traffic conditions. The service time of node A is determined by the 20-minute loading operation time, the service time of node B is determined by the 20-minute loading operation time, the service time of node C is determined by the 15-minute unloading operation time, and the service time of node D is determined by the 15-minute unloading operation time. Using an improved cost-saving algorithm, under the premise of satisfying vehicle load and volume constraints, a Hamiltonian path is constructed in the access graph model, starting from the virtual starting point S, visiting all nodes, and finally reaching the virtual ending point E. The initially constructed Hamiltonian path is S->A->B->C->D->E.

[0039] The constructed Hamiltonian path is locally optimized by swapping the visit order of adjacent nodes or adjusting sub-paths to reduce the total travel time. The initial Hamiltonian path S->A->B->C->D->E, constructed by the path search algorithm, is obtained. A two-point swapping strategy is employed, sequentially attempting to swap the visit order of any two internal nodes that are neither the starting nor ending point of the path. The visit order of nodes B and C is then swapped to form a new visit sequence S->A->C->B->D->E. For the newly generated visit sequence S->A->C->B->D->E, its total travel time under the current real-time traffic conditions is recalculated. Total travel time: in: This indicates the total travel time for the access sequence; This represents the complete path formed by the order in which nodes are visited. This represents a directed edge on a path from node m to the next node n; This represents the travel time required to travel from node m to node n under real-time traffic conditions. This represents the set of all nodes on the path that require service. This represents the service time for loading or unloading operations at node k. The total travel time for the new access sequence is calculated. The total travel time for the original Hamilton route is 220 minutes. The total travel time of the new access sequence is 240 minutes. The total travel time of the new access sequence is less than that of the original Hamiltonian path, and it has been verified that it does not violate the expected time period constraints of each logistics delivery request. The original path is replaced with the new sequence S->A->C->B->D->E.

[0040] After completing all two-point swap attempts, a path segment reversal strategy is employed. A consecutive node segment is randomly selected from the current path S->A->C->B->D->E, for example, nodes A, C, and B are chosen as the path segment. The visiting order of the path segment A->C->B is reversed, resulting in B->C->A, forming a new visiting sequence S->B->C->A->D->E. The total travel time of the new visiting sequence S->B->C->A->D->E is evaluated to see if it is reduced, and the time is recalculated. The total travel time of the new access sequence is 210 minutes. Since the total travel time of the current path is 220 minutes, this change is accepted, and the path is updated to S->B->C->A->D->E. The path segment reversal process is repeated; for example, nodes C, A, and D are selected as segments for reversal. If the total travel time does not decrease after evaluation, this change is abandoned, and this process is repeated until the total travel time cannot be further optimized. The final optimized access sequence is output as S->B->C->A->D->E.

[0041] In some embodiments, the improved saving algorithm considers both the saving of travel time and the satisfaction of time window constraints when calculating the saving value, prioritizing the merging of nodes with high saving values ​​that do not violate the time window. In some embodiments, in addition to two-point swaps and path segment reversals, local optimization may also employ a node insertion strategy, removing a node from its current position and inserting it into other possible positions on the path to evaluate changes in total travel time. Optionally, when evaluating whether a new access sequence violates the expected time period constraint, a fast comparison is made between the pre-calculated actual arrival time of each node under the candidate sequence and the expected time window. Optionally, when goods contain multiple receiving locations, the virtual endpoint can be the last received location visited or a virtual node returning to the vehicle distribution center. It can be understood that local optimization seeks access sequences with shorter total travel times by iteratively adjusting the node order. The construction and optimization process of the Hamiltonian path aims to efficiently obtain high-quality feasible access sequences in large-scale node searches.

[0042] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for intelligent group buying of multi-source logistics demand along the same route based on vehicle positioning, characterized in that: Includes the following steps: Receive logistics shipment requests from multiple different shippers. The logistics shipment requests include the shipping location, receiving location, cargo specifications, and expected time period information. The system can obtain the real-time location and attribute information of available freight vehicles within the area. The vehicle attribute information includes cargo space and rated load capacity. Based on the shipping and receiving locations of all logistics delivery requests, many-to-many route similarity calculations are performed to identify request combinations with geographical overlap. The identified request combinations of cargo specifications are aggregated and matched with the obtained vehicle attribute information to filter out the target vehicle set that carries the aggregated cargo. Based on the real-time location of each vehicle in the target vehicle set, calculate the estimated travel time to each delivery location in each request combination; By combining the estimated travel time with the expected time period of each logistics delivery request, at least one feasible multi-source group-buying transportation plan is generated for each request. The multi-source group-buying transportation plan includes specific vehicles, picking order and time nodes.

2. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 1, characterized in that, The process involves calculating many-to-many route similarity based on the shipping and receiving locations of all logistics delivery requests to identify request combinations with geographical overlap. Specifically, this includes: Map the shipping and receiving locations of all logistics shipment requests to a sequence of geographic coordinates, with each request forming a transportation line segment pointing from the shipping coordinates to the receiving coordinates; Calculate the straight-line distance between the starting points and the ending points of any two transport segments, as well as the average distance between the two segments in space; The calculated straight-line distance and average distance are input into the preset route overlap assessment model, which outputs a similarity score by combining the distance and the direction angle. Two logistics delivery requests with similarity scores higher than a set overlap threshold are grouped into the same set to be evaluated. A set to be evaluated may contain more than two requests. Based on the set to be evaluated, traverse all the shipping and receiving locations of all requests within the set to be evaluated, calculate the smallest convex polygon region covering all locations in the set to be evaluated, and mark the smallest convex polygon region as the collinear region of the request combination.

3. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 2, characterized in that, The process of aggregating the identified request combinations of cargo specifications and matching them with the acquired vehicle attribute information specifically includes: Extract the cargo specifications of all logistics shipping requests in a request combination. The cargo specifications include volume, weight and cargo type. The volumes of all goods in the requested combination are summed to obtain the total aggregate volume, and the weights of all goods are summed to obtain the total aggregate weight. Based on the type of goods, determine whether there are loading constraints, including the prohibition of mixing special goods, and perform virtual loading simulation on aggregated goods based on the judgment result; Iterate through the real-time acquired vehicle attribute information to filter out candidate vehicles whose cargo space is greater than the aggregated total volume and whose rated load is greater than the aggregated total weight. For the selected candidate vehicles, a second verification is performed based on their cargo box structure parameters and the results of virtual loading simulation. Vehicles that cannot meet the cargo type loading constraints are eliminated, and the remaining vehicles constitute the target vehicle set.

4. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 3, characterized in that, The step of calculating the estimated travel time to each delivery location in each request combination based on the real-time location of each vehicle in the target vehicle set specifically includes: Obtain the real-time latitude and longitude coordinates of each vehicle in the target vehicle set, and request the geographic coordinates of all delivery locations in the combined collinear area; Based on real-time traffic data, the optimal driving route from the real-time location of each vehicle to each delivery location is planned, taking into account the current road congestion. Based on the length of the planned optimal driving route and the real-time road traffic speed, calculate the travel time for each vehicle to reach each delivery location; Based on the travel time en route, a fixed vehicle preparation time and a loading operation baseline time are added to obtain the total estimated travel time from the vehicle's current location to the point where loading is completed. Construct an estimated travel time matrix, where the rows of the matrix correspond to the vehicles in the target vehicle set, the columns correspond to the delivery locations in the request combination, and the matrix elements are the corresponding total estimated travel time.

5. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 4, characterized in that, The process combines the estimated travel time with the expected time slots for each logistics delivery request to generate at least one feasible multi-source group-buying transportation plan for each request, specifically including: For a request combination and its corresponding estimated time matrix, each row of the matrix, i.e. each vehicle, is used as the vehicle to execute the group-buying transportation plan. Starting with the vehicle to be executed, based on the time consumption data in the estimated time consumption matrix, a path search algorithm is used to find a sequence that visits all delivery locations of the requested combination and finally completes all deliveries. The sequence must satisfy that each location is visited only once. While searching the access sequence, calculate the actual arrival time of each shipping location in the sequence. The actual arrival time is equal to the departure time of the previous location in the sequence plus the travel time from the previous location to the current location. The actual arrival time of each shipping location is calculated and compared with the expected time period information of the logistics shipping request corresponding to the shipping location to check whether it falls within the expected time period. If there exists an access sequence such that the actual arrival time of all shipping locations meets their respective expected time periods, then the executing vehicle is bound to the access sequence to generate a complete group-buying transportation plan, in which the estimated operation time of each node is specified.

6. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 5, characterized in that, The method employs a path search algorithm to find a sequence that combines all shipping locations in the request and ultimately completes all deliveries. Specifically, this includes: Set the search starting point as the real-time location of the vehicle being executed, and treat all shipping and receiving locations in the request combination as the set of nodes that must be visited. In the node set, the real-time location of the executing vehicle is introduced as a virtual starting point, and the last receiving location is introduced as a virtual ending point to construct a complete access graph model. In the access graph model, the weight of an edge is determined by the travel time between two points under real-time traffic conditions, and the service time of a node is determined by the loading and unloading time of goods. Using an improved saving or insertion algorithm, under the premise of satisfying vehicle load and volume constraints, a Hamiltonian path is constructed in the access graph model that starts from a virtual starting point, visits all nodes, and finally reaches a virtual ending point. The constructed Hamiltonian path is locally optimized by swapping the visit order of adjacent nodes or adjusting sub-paths to reduce the total travel time, and finally the optimized visit sequence is output.

7. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 6, characterized in that, The calculation covers the smallest convex polygon region that encompasses all locations in the set to be evaluated, specifically including: Read the geographic coordinates of the shipping and receiving locations of all logistics shipping requests in a set to be evaluated, and form a set of geographic coordinate points to be covered. Find the point with the smallest latitude from the set of geographic coordinates. If there are multiple points, select the point with the smallest longitude as the starting point. Using the starting base point as the origin of polar coordinates, calculate the polar angles of all other points in the point set relative to the origin of polar coordinates, and sort them in order of increasing polar angles; Process each sorted point in turn, compare the current point with the last two points that have formed the boundary of the convex polygon, and determine whether it forms a convex edge or needs to backtrack to delete the concave point. After traversing all points, connect the finally determined points to form a convex polygon, which is the smallest convex polygon region covering all the points of the requested combination, and record the vertex coordinates of the smallest convex polygon.

8. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 7, characterized in that, The process of planning the optimal driving route from the real-time location of each vehicle to each delivery location based on real-time traffic data specifically includes: Connect to the city's real-time traffic information service platform to obtain current road network topology data, average driving speed of each road segment, and congestion event information; The road network is modeled as a weighted directed graph, where nodes are road intersections and the weights of edges represent the estimated travel time of a road segment under the current road conditions. In a weighted directed graph, the real-time location of a vehicle is mapped to the nearest road node as the starting point of the path, and the delivery location is mapped to the nearest road node as the ending point of the path. Run the shortest path algorithm on a weighted directed graph to find the sequence of nodes with the minimum cumulative weight from the starting node to the ending node, where the weight represents the shortest travel time. The node sequence found by the shortest path algorithm is mapped back to the actual road to generate a specific optimal driving path that takes into account real-time traffic conditions, and its total length and estimated travel time are recorded.

9. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 8, characterized in that, The actual arrival time of each shipping location in the calculation sequence specifically includes: Determine the start time of the group-buying transportation plan, where the start time is the initial moment when the vehicle departs from its real-time location; Read the generated access sequence. The first element of the sequence is the first delivery location to be accessed. Obtain the total estimated travel time from the real-time vehicle location to the first delivery location from the estimated travel time matrix. Add the initial time to the total estimated travel time to obtain the actual arrival time of the vehicle at the first delivery location; After obtaining the actual arrival time at the first shipping location, add the standard cargo loading operation time at that shipping location to obtain the time when the vehicle leaves the shipping location; Based on the time the vehicle leaves the previous position, and adding the travel time from the previous position to the next position in the sequence obtained from the estimated time matrix, the actual arrival time of the next position in the sequence is iteratively calculated until all positions in the sequence have been calculated.

10. The intelligent group-buying method for multi-source logistics demand along the same route based on vehicle positioning according to claim 9, characterized in that, The local optimization of the constructed Hamiltonian path, which reduces the total travel time by swapping the visit order of adjacent nodes or adjusting sub-paths, specifically includes: Obtain the Hamiltonian path initially constructed by the path search algorithm, which includes the access order of all nodes from the virtual starting point to the virtual ending point; A two-point swapping strategy is adopted to try swapping the access order of any two internal nodes that are neither the starting point nor the ending point in the path in turn, forming a new access sequence; For each newly generated access sequence, recalculate its total travel time under the current real-time traffic conditions, including the travel time between nodes and the service time on the nodes; If the total travel time of the new access sequence is less than the total travel time of the original Hamiltonian path, and does not violate the expected time period constraints of each logistics delivery request, then the new sequence replaces the original path; After completing all two-point swap attempts, a path segment reversal strategy is adopted. A continuous node segment in the path is randomly selected, and the access order of the path segment is reversed. The total travel time is evaluated to see if it is reduced. If it is reduced, the change is accepted. This process is repeated until the total travel time cannot be further optimized.

Images