Truck-unmanned aerial vehicle cooperative path planning method and system

Abstract

The application discloses a truck-unmanned aerial vehicle cooperative path planning method and system. The method comprises the following steps: inputting parameter information of a vehicle path problem with an unmanned aerial vehicle considering an unmanned aerial vehicle exchange mode, calling a heuristic algorithm, and outputting an initial upper bound solution scheme; inputting the initial upper bound solution scheme, calling a column generation algorithm, solving a subproblem in the column generation algorithm by using a cooperative path label algorithm, finding a path column by using a heuristic algorithm, and outputting a linear solution scheme; inputting the linear solution scheme, calling a branching strategy and a cut constraint search strategy, adding a branch and a cut constraint after solving the column generation algorithm at each branching node, and outputting an integer solution scheme. According to the scheme, the unmanned aerial vehicle can be exchanged between different trucks without relying on an unmanned aerial vehicle parking point, and the solving efficiency of an accurate algorithm is improved.

Description

Technical Field

[0001] The present invention generally relates to the field of path planning technology. More specifically, the present invention relates to a truck-UAV collaborative path planning method and system. Background Art

[0002] As a new mode of transportation for logistics and delivery, drones have attracted widespread attention from logistics companies due to their high speed and low cost. However, due to their limited range and payload capacity, drones cannot independently carry out large-scale logistics and delivery operations. Instead, they require coordinated delivery with other transport vehicles such as trucks.

[0003] The main problems in the collaborative delivery of trucks and drones are the traveling salesman problem with drones (TSPD) and the vehicle routing problem with drones (VRPD). In 2019, Wang and Sheu proposed a branch and price algorithm for the VRPD problem in the drone exchange and recycling scenario. Figure 1As shown in Figure 2, the VRPD problem involves multiple trucks, each carrying multiple drones, departing from a depot and returning to the depot after completing service at customer points in the network. Their proposed scenario introduces a new type of drone docking hub, which is equivalent to connecting drones between trucks through the docking hub. In this scenario, a pulse algorithm is used as a subproblem of the column generation algorithm. However, in the BP algorithm designed by Wang and Sheu (2019) for the VRPD problem considering drone docking hubs, drone recovery is completely linked to the docking hub, which is equivalent to severing the connections between different trucks through the docking hub. The main impact is that it reduces the overall scale of the problem and the feasible solution space, thus preventing some more optimal drone exchange solutions from being fully considered. In real life, the establishment of additional drone docking hubs constitutes a fixed equipment investment, and the location of drone docking hubs is another difficult issue that needs to be considered. On the other hand, the exact algorithm frameworks for the truck and drone collaborative delivery problem have relatively low solution efficiency. The BP algorithm proposed by Roberti and Ruthmair (2021) for TSP-D can only solve a maximum of 39 customer points. The BP algorithm designed by Wang and Sheu (2019) for the VRPD problem with drone parking points can only solve a maximum of 15 customer points. Kitjacharoenchai et al. (2019) solved a large-scale VRPD problem with drone interchange mode, but used a heuristic algorithm that could not guarantee the quality and accuracy of the solution. In summary, how to enable drones to interchange with different trucks independently of the drone parking point and speed up the solution efficiency of the exact algorithm has become an urgent problem to be solved. Summary of the Invention

[0004] To at least address the technical issues described in the background technology section above, the present invention proposes a truck-UAV collaborative path planning method, system, and storage medium. Utilizing this solution, by introducing three special processing methods: pioneering customer points, path groups, and information sharing mechanisms, a precise branching, pricing, and pruning (BPC) algorithm framework for the VRPD problem is designed to account for UAV interchangeability. This allows UAVs to interchange with different trucks independently of their parking locations. Furthermore, by incorporating a series of effective acceleration strategies into the BPC precise algorithm framework, the precise algorithm's solution efficiency is accelerated. In light of this, the present invention provides solutions in the following aspects.

[0005] The first aspect of the present invention provides a truck-UAV collaborative path planning method, including: S1, inputting parameter information of a vehicle routing problem with UAVs (VRPD) considering a UAV interchange mode, calling a heuristic algorithm, and outputting an initial upper bound solution; S2, inputting the initial upper bound solution, calling a column generation algorithm, using a collaborative path labeling algorithm to solve subproblems in the column generation algorithm, using a heuristic algorithm to assist in finding a path column, and outputting a linear solution; S3, inputting a linear solution, calling a branching strategy and a cut constraint search strategy, adding branches and subset row inequalities after solving each branch node using the column generation algorithm, and outputting an integer solution.

[0006] In one embodiment, the collaborative path labeling algorithm includes: S21, input label L i The information of other paths in the same path group is obtained through the information sharing mechanism, and the Ng-route relaxation strategy is called to output the label L i The customer point set V that can be expanded in the future suit ; S22, input L i and V suit , L i V suit Each customer point j in the tag expansion rule is called to generate a new tag L j The set of; for the unruled label L j , enumerate L j In the case of drone release and recovery, the pioneer customer point set CFSset ​​and the new label L are output j Set, add the new label to the label bucket; S23, input L i and CFSset,L i Call the label expansion rule for each pioneer customer point j′ in CFSset ​​to generate a new label L temp The set of; if L temp and L j′ If the paths are different, then L temp and L j′ Add to the same path group G temp ; In the generation of L temp After that, the path group G needs to be updated through the information sharing mechanism. temp Information; output new label L temp The collection adds the new tag to the tag bucket.

[0007] In one embodiment, obtaining information about other paths within the same path group through an information sharing mechanism includes: immediately updating the information contained in the path group using the path group shared information as a medium after a new path group is constructed or path information is changed; and obtaining information about other paths through the path group shared information when other paths within the path group are expanded, thereby identifying a set of customers that can be subsequently expanded.

[0008] In one embodiment, the pioneer customer point is specifically: when drone u returns to truck customer point j after serving a drone customer point i, before a truck k arrives at customer point j, customer point j is called the pioneer customer point associated with drone u.

[0009] In one embodiment, the L temp and L j′ Add to the same path group G temp , including four situations: non-path group path constructs path group, non-path group path is associated with path group, path group path is associated with non-path group path and path group path is associated with path group: the non-path group path constructs path group, which is to summarize the information of different truck paths, extract the information required by the path group, and construct a new path group; the non-path group path is associated with the path group and the path group path is associated with the non-path group path, both need to add the new path to the path group and update the path group information according to the newly added path; the path group path is associated with the path group, which is to integrate the information of the two merged path groups, remove duplicates and other operations, and obtain a new merged path group.

[0010] In one embodiment, the mutual dominance between the two path groups requires that when all paths in one path group can dominate all paths in the other path group, the two path groups can dominate each other.

[0011] In one embodiment, the acceleration is also performed by applying a two-dimensional bucketing strategy, wherein the two-dimensional bucketing strategy includes: dividing all tags extended to the same customer point into the same bucket; and further dividing all tags extended to the same customer point into smaller buckets according to the cumulative load.

[0012] In one embodiment, the acceleration is further performed by applying an Ng-route relaxation strategy, wherein the Ng-route relaxation strategy includes: when a path in a path group generates a loop, the truck path generation loop can only repeatedly visit customer points that have been served by the current path, and cannot repeatedly visit customer points that have been served by other truck paths in the path group.

[0013] In one embodiment, it also includes applying heuristic-assisted column pool management for acceleration, and the heuristic-assisted column pool management includes: constructing a column pool to store path column solutions, and the column pool is filled by multiple heuristic algorithms and collaborative path labeling algorithms; in each iteration of the column generation algorithm, first check whether there are path columns with negative estimated costs in the column pool, if so, select one or more path columns with the smallest estimated costs and add them to the main problem; at the same time, limit the maximum size of the column pool. If it exceeds this size, then keep a part of the path columns with the smallest path cost in all path columns in the column pool, and delete the remaining path columns in the column pool.

[0014] A second aspect of the present invention provides a truck-UAV collaborative path planning system, which applies any of the truck-UAV collaborative path planning methods described above.

[0015] In response to the shortcomings of the existing technology, the present invention introduces three special processing methods: pioneer customer points, path groups, and information sharing mechanisms. A collaborative path labeling algorithm for the VRPD problem considering the drone exchange mode is designed and implemented in the Branch Price and Cuts (BPC) precise algorithm framework, so that drones can be exchanged between different trucks without relying on the drone parking points. For the collaborative path problem, a series of effective acceleration strategies are added to the BPC precise algorithm framework to speed up the solution efficiency of the precise algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an illustrative and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0017] Figure 1 It is to show the branch pricing algorithm in the prior art;

[0018] Figure 2 1. It illustrates a truck-UAV collaborative path planning method according to an embodiment of the present invention;

[0019] Figure 3 FIG. 2 is a diagram illustrating a cooperative path numbering algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0021] It should be understood that the terms "first," "second," "third," and "fourth," etc. in the claims, description, and drawings of the present invention are used to distinguish different objects, rather than to describe a specific order. The terms "comprise" and "comprising" used in the description and claims of the present invention indicate the presence of the described features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or combinations thereof.

[0022] It should also be understood that the terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the present invention. As used in the specification and claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should further be understood that the term "and / or" as used in the specification and claims refers to any and all possible combinations of one or more of the associated listed items, including and including these combinations.

[0023] As used in this specification and claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0024] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] A first aspect of the present invention provides a truck-UAV collaborative path planning method. Figure 2 The method for truck-UAV collaborative path planning according to an embodiment of the present invention is shown, which can be described as including steps S1-S3:

[0026] S1. Input the parameter information of the vehicle routing problem with drones (VRPD) considering the drone interchange mode, call the heuristic algorithm, and output the initial upper bound solution;

[0027] S2. Input the initial upper bound solution, call the column generation algorithm, use the collaborative path labeling algorithm to solve the subproblems in the column generation algorithm, use the heuristic algorithm to assist in finding the path column, and output the linear solution;

[0028] S3. Input the linear solution, call the branching strategy and cut constraint search strategy, add branches and subset row inequalities after solving each branch node using the column generation algorithm, and output the integer solution.

[0029] As you can understand, drone-truck collaboration generally has two modes: interchange mode and remote mode. The interchange mode means a drone on a truck k1 is retrieved by another truck k2, while the remote mode means a drone released from the same truck is retrieved by the same truck.

[0030] Furthermore, in an embodiment of the present invention, under the above-mentioned truck-UAV collaborative path planning actuarial algorithm framework for the VRPD problem considering the UAV interchange mode, a collaborative path labeling algorithm is adopted in the process of solving the column generation algorithm sub-problem for the collaborative path problem.

[0031] In a preferred embodiment of the present invention, a collaborative path labeling algorithm process is used to solve the VRPD problem considering the UAV interchange mode by utilizing three special processing means: pioneer client points, path groups, and information sharing mechanisms.

[0032] In this invention, because each label in the labeling algorithm can only represent a continuous path, label expansion using this algorithm cannot account for drones being exchanged and retrieved between different trucks. Similar issues exist for other collaborative path problems. To overcome this bottleneck, the invention introduces pioneering customer points: when a drone path r is expanded to a truck customer point i that has not been visited by the current label, the drone path is "virtually" retrieved at customer point i. This means that a copy of customer point i is created, i', which contains information about drone path r. Subsequently, when the truck path is expanded, both the set of regular customer points and the set of pioneering customer points are expanded simultaneously. When expansion reaches a pioneering customer point i', the truck and drone convergence process is complete. This process enables drones to be exchanged and retrieved between different trucks. A pioneering customer point is a customer point associated with drone u when drone u returns to truck customer point j after serving drone customer point i and before truck k reaches customer point j.

[0033] In the collaborative path labeling algorithm, when there are unreleased drones on a truck on a newly expanded path, it is necessary to enumerate the release and recovery situations of the remaining drones on the truck, mark the drone recovery points as pioneer customer points, and add them to the pioneer customer point set; thus, in the subsequent truck path expansion process, the truck can expand to the pioneer customer points of the sub-path carrying the drone to perform drone recovery operations.

[0034] In the present invention, when two or more truck paths are interconnected due to drone exchange, these truck paths need to share information, including all drone paths associated with these truck paths, to form a path group. The path group refers to when multiple truck paths k1, k2, ..., k m After the drone exchange recovery operation, the truck paths k1, k2, ..., k m They need to be merged into a path group to facilitate information acquisition and label expansion in subsequent labeling algorithms. The information contained in each path group is as follows:

[0035] ① The set of all truck customer points and drone customer points visited by the path group;

[0036] ② The union of the set of customer points visited by all truck / UAV paths in the path group and the set of customer points that cannot be visited;

[0037] ③ The cumulative cost of all truck routes in the route group;

[0038] ④ The cumulative number of drones carried by all trucks in the path group;

[0039] ⑤ The association relationship between all truck routes in the route group and the route group, that is, recording which truck routes the current route group contains.

[0040] Since the path group ultimately forms a set of feasible paths, after the path group is formed, when the paths in the path group are subsequently expanded, the set of customer points that have been visited in the path group and the set of customer points that cannot be visited by this path need to be used as criteria to determine the set of customer points that can be visited subsequently.

[0041] The construction of a route group includes the following four cases: non-route group paths building a route group, non-route group paths being associated with a route group, route group paths being associated with non-route group paths, and route group paths being associated with a route group. In the case of non-route group paths building a route group, the information of different truck routes is aggregated, the information required for the route group is extracted, and a new route group is constructed. In the case of non-route group paths being associated with a route group and route group paths being associated with non-route group paths, the new route needs to be added to the route group and the route group information is updated based on the newly added route. In the case of route group paths being associated with a route group, the information of the two merged route groups is integrated, and after deduplication and other operations, a new merged route group is obtained.

[0042] In the collaborative path labeling algorithm, when different truck paths need to be associated with each other after a drone exchange and recovery operation, it is necessary to construct a path group for the different truck paths, that is, to put the different truck paths into the same path set; the different truck paths in the path group need to be able to obtain information about other paths, so as to ensure the accuracy of the solution obtained by the paths in the path group during the subsequent expansion process.

[0043] In the present invention, an information sharing mechanism is also used to ensure the accuracy of the solution. That is, all paths in the same path group need to share the information contained in the path group. In other words, any path in the path group can obtain the path information contained in other paths in the path group, ensuring the accuracy of the path solution in the same path group.

[0044] The above information sharing mechanism is implemented through the following scheme: when a new path group is constructed or the information of any path in the path group changes due to path expansion, it is necessary to use the path group shared information as a medium to update the information contained in the path group immediately after the path group is constructed or the path information is changed; then, when other paths in the path group are expanded, the information of other paths can be obtained through the path group shared information, thereby finding a set of customers that can be expanded subsequently.

[0045] In the collaborative path labeling algorithm, the paths in the path group need to use the information sharing mechanism before and after expansion. i Before expansion, it is necessary to obtain information about other paths through the shared information of other paths in the path group, so as to facilitate the subsequent ng-route relaxation strategy to determine the current path L i The customer set that can be expanded later; a path L in the path group i After the expansion, the path information in the entire path group needs to be updated through the information sharing mechanism to facilitate the acquisition of this information when other paths are expanded later.

[0046] In the embodiment of the present invention, for any label Li , using the collaborative path labeling algorithm to expand it from customer point i to customer point j. The whole expansion steps are as follows Figure 3 As shown, it can be described as including steps S21-S23:

[0047] S21, input label L i The information of other paths in the same path group is obtained through the information sharing mechanism, and the Ng-route relaxation strategy is called to output the label L i The customer point set V that can be expanded in the future suit ;

[0048] S22, input L i and V suit , L i V suit Each customer point j in the tag expansion rule is called to generate a new tag L j The set of; for the unruled label L j , enumerate L j In the case of drone release and recovery, the pioneer customer point set CFSset ​​and the new label L are output j Add the new tag to the tag bucket;

[0049] S23, input L i and CFSset,L i Call the label expansion rule for each pioneer customer point j′ in CFSset ​​to generate a new label L temp The set of; if L temp and L j′ If the paths are different, then L temp and L j′ Add to the same path group G temp ; In the generation of L temp After that, the path group G needs to be updated through the information sharing mechanism. temp Information; output new label L temp The collection adds the new tag to the tag bucket.

[0050] The above-mentioned label expansion step shows the single expansion process of a certain label during the operation of the collaborative path labeling algorithm. When all labels are expanded to the site after multiple expansions, the collaborative path labeling algorithm ends. The process from the initial label (the beginning of the collaborative path labeling algorithm) to the expansion of all labels to the site (the end of the collaborative path labeling algorithm) is called a collaborative path labeling algorithm. At the same time, in the above process, the label expansion rules and domination rules are similar to the BP algorithm for the TSP-D problem. At the same time, this technical solution stipulates that the mutual domination between two path groups requires that all paths in one path group can dominate all paths in the other path group. In this case, the two path groups can dominate each other.

[0051] In a preferred embodiment of the present invention, a series of acceleration strategies are added to the algorithm framework to speed up the algorithm's solution efficiency while ensuring the accuracy of the problem solution:

[0052] Two-dimensional bucketing strategy: This strategy is designed based on two dimensions: the customer points to which a tag is extended and the cumulative payload of the tag. In the first dimension, all tags extended to the same customer point are grouped into the same bucket. Then, based on the second dimension (cumulative payload), all tags extended to the same customer point are further divided into smaller buckets. This accelerates subsequent rule determination and improves the algorithm's solution efficiency.

[0053] Ng-route relaxation strategy: Ng-route relaxation relaxes the restriction that each customer point in the subproblem can only be visited once by allowing the pricing subproblem to generate a loop path, thereby simplifying the difficulty of solving the subproblem. The path types in the present invention include truck paths and drone subpaths. Therefore, the limiting conditions when applying Ng-route relaxation are: since a drone can only serve one customer point, it is first stipulated that the customer point served by the drone subpath cannot be used to generate a loop path, that is, the customer point served by the drone does not participate in the loop construction process of Ng-route relaxation and is directly added to the set of customer points that cannot form a loop; secondly, when a customer point served by a truck can generate a loop path after being determined by Ng-route relaxation, when the truck subsequently repeatedly visits this customer point, the release and recovery processes of the drone at this repeatedly visited customer point are not considered. Because the final solution path definitely does not contain loops, trucks cannot repeatedly visit customer points to release drones. Considering that releasing and recovering drones significantly increases computational complexity (increasing the number of precursor customer points, labels, and paths), no operations related to drone sub-paths are performed after loops are generated. Because the drone interchange pattern leads to operations related to path groups, when generating loops, this paper only operates on individual truck paths, not on path groups. That is, when generating loops for a truck path, it can only repeatedly visit customer points already served by the current path, and cannot repeatedly visit customer points already served by other truck paths in the path group.

[0054] Heuristic-assisted column pool management: First, a column pool is constructed to store path list solutions. The column pool is populated using a series of heuristic algorithms and a collaborative path labeling algorithm. A multi-entry approach is also employed: During each iteration of the column generation algorithm, the column pool is checked for any path lists with negative cost. If so, one or more paths with the lowest cost are selected and added to the main problem. A maximum column pool size is also set. If this size is exceeded, the path lists with the lowest cost are retained in the column pool, and the remaining path lists are deleted.

[0055] In order to reduce the number of branches in the algorithm solving process, the present invention adopts the method of adding effective subset row inequalities (Sub-row Inequalities) to reduce the number of branches and improve the algorithm solving efficiency for the VRPD problem considering the UAV interchange mode.

[0056] based on Figure 2-Figure 3The described truck-UAV collaborative path planning method, the second aspect of the present invention also provides a truck-UAV collaborative path planning method system, the above-mentioned truck-UAV collaborative path planning method system runs the above-mentioned truck-UAV collaborative path planning method.

[0057] Although this specification has shown and described a plurality of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will appreciate that many modifications, variations, and alternatives can be made without departing from the spirit and scope of the present invention. It should be understood that in practicing the present invention, various alternatives to the embodiments of the present invention described herein may be employed. The appended claims are intended to define the scope of protection of the present invention and therefore cover modular compositions, equivalents, or alternatives within the scope of these claims.

Claims

1. A truck-UAV collaborative path planning method, characterized in that: include; S1. Input the parameter information of the vehicle routing problem with drones (VRPD) considering the drone interchange mode, call the heuristic algorithm, and output the initial upper bound solution; S2. Input the initial upper bound solution, call the column generation algorithm, use the collaborative path labeling algorithm to solve the subproblems in the column generation algorithm, use the heuristic algorithm to assist in finding the path column, and output the linear solution; S3. Input the linear solution, call the branching strategy and the cut constraint search strategy, add branches and subset row inequalities after solving each branch node using the column generation algorithm, and output the integer solution; The collaborative path labeling algorithm includes: S21, input label L i The information of other paths in the same path group is obtained through the information sharing mechanism, and the Ng-route relaxation strategy is called to output the label L i The customer point set V that can be expanded in the future suit ; S22, input L i and V suit , L i V suit Each customer point j in the tag expansion rule is called to generate a new tag L j The set of; for the unruled label L j , enumerate L j In the case of drone release and recovery, the pioneer customer point set CFSset ​​and the new label L are output j Add the new tag to the tag bucket; S23, input L i and CFSset,L i Call the label expansion rule for each pioneer customer point j′ in CFSset ​​to generate a new label L temp The set of; if L temp and L j′ If the paths are different, then L temp and L j′ Add to the same path group G temp ; In the generation of L temp After that, the path group G needs to be updated through the information sharing mechanism. temp Output new label L temp Add the new tag to the tag bucket; The pioneer customer point is specifically: when drone u returns to truck customer point j after serving a drone customer point i, before a truck k arrives at customer point j, customer point j is called the pioneer customer point associated with drone u.

2. The truck-UAV collaborative path planning method according to claim 1, characterized in that: The information of other paths in the same path group obtained through the information sharing mechanism includes: When a new path group is constructed or the path information is changed, the path group shared information is used as a medium to update the information contained in the path group; When other paths in the path group are expanded, information of other paths is obtained through the path group shared information, so as to find a customer set that can be expanded subsequently.

3. The truck-UAV collaborative path planning method according to claim 1, characterized in that: The L temp and L j′ Add to the same path group G temp , including four situations: non-path group path builds path group, non-path group path is associated with path group, path group path is associated with non-path group path and path group path is associated with path group: The non-path group path construction path group is to aggregate the information of different truck paths, extract the information required for the path group, and construct a new path group; When the non-path group path is associated with the path group and the path group path is associated with the non-path group path, it is necessary to add the new path to the path group and update the path group information according to the newly added path; The path group path is associated with the path group by integrating the information of the two merged path groups and performing a deduplication operation to obtain a new merged path group.

4. According to a truck-UAV collaborative path planning method according to any one of claims 1 to 3, the mutual dominance between the two path groups requires that when all paths in one path group can dominate all paths in the other path group, the two path groups can dominate each other.

5. The truck-UAV collaborative path planning method according to claim 1, further comprising applying a two-dimensional bucketing strategy for acceleration, wherein the two-dimensional bucketing strategy comprises: All tags extended to the same customer point are divided into the same bucket; All tags extending to the same customer point are further subdivided into smaller buckets based on the cumulative payload.

6. The truck-UAV collaborative path planning method according to claim 1 further comprises applying an Ng-route relaxation strategy for acceleration, wherein the Ng-route relaxation strategy comprises: When a path in a path group generates a loop, the truck path generation loop can only repeatedly visit customer points that have been served by the current path, and cannot repeatedly visit customer points that have been served by other truck paths in the path group.

7. The truck-UAV collaborative path planning method according to claim 1, further comprising applying heuristic-assisted column pool management for acceleration, wherein the heuristic-assisted column pool management comprises: A column pool is constructed to store path column solutions. The column pool is filled by multiple heuristic algorithms and collaborative path labeling algorithms. In each iteration of the column generation algorithm, the column pool is first checked to see if there are path columns with negative estimated costs. If so, one or more path columns with the smallest estimated costs are selected and added to the main problem. At the same time, the maximum size of the column pool is limited. If this size is exceeded, a portion of the path columns with the smallest path costs among all path columns are kept in the column pool, and the remaining path columns in the column pool are deleted.

8. A truck-UAV collaborative path planning system, characterized in that: Apply the truck-UAV collaborative path planning method as described in any one of claims 1 to 7.

Images