Vehicle-cabinet collaborative distribution path optimization method and system

Abstract

The present invention provides a vehicle-machine cabinet collaborative distribution path optimization method and system, which relates to the field of path optimization technology. The present invention first obtains collaborative distribution task data with coverage points; then, based on the collaborative distribution task data, a vehicle-machine collaborative distribution model with a two-level path with coverage points is constructed with the goal of minimizing the total distribution cost; finally, the vehicle-machine collaborative distribution model is solved through a two-stage heuristic algorithm to obtain an optimized path. The vehicle-machine collaborative distribution model with a two-level path with coverage points constructed in the present invention can describe the vehicle-machine collaborative distribution problem in the two-level path problem while considering the introduction of the coverage point option, which is more in line with the actual process of vehicle-machine collaborative last-mile distribution. By solving the model, the optimized path obtained can effectively improve distribution efficiency and reduce terminal distribution costs.

Description

Technical Field

[0001] The present invention relates to the technical field of route optimization, and in particular to a method and system for optimizing the route of vehicle-cabinet collaborative distribution. Background Art

[0002] As drone technology matures, more and more logistics companies are beginning to use drones in express delivery. The vehicle routing problem of collaborative delivery between drones and trucks has become a hot topic in academia and industry.

[0003] At present, domestic and foreign scholars' research on this issue mainly focuses on the modeling and solution algorithms of the traveling salesman problem with drones, the vehicle routing problem with drones and their derivative problems.

[0004] However, in actual last-mile delivery, due to the existence of coverage points such as express lockers and Cainiao stations, the solved delivery path does not match the actual situation, resulting in low delivery efficiency. Summary of the Invention

[0005] (1) Technical problems solved

[0006] In view of the shortcomings of the existing technology, the present invention provides a vehicle-cabinet collaborative distribution path optimization method and system, which solves the technical problem of low distribution efficiency of the existing method.

[0007] (2) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] In a first aspect, the present invention provides a method for optimizing a vehicle-cabinet collaborative delivery path, the method comprising:

[0010] S1. Obtain collaborative delivery task data with coverage points;

[0011] S2. Based on the collaborative delivery task data, a vehicle-machine collaborative delivery model with a two-level path and coverage points is constructed with the goal of minimizing the total delivery cost.

[0012] S3. Solve the vehicle-machine collaborative delivery model through a two-stage heuristic algorithm to obtain the optimized path.

[0013] Preferably, the collaborative delivery task data with coverage points includes: a delivery network, a transport tool set, a node set, and node information;

[0014] The distribution network refers to the directed graph formed between warehouses, drone sites, coverage points, and customer locations, represented by G = (V, A), where V represents the set of all nodes. A represents the set of all arcs, that is, the set of all delivery paths, A = {(i, j) | i, j∈V, i≠j};

[0015] The transportation tool set includes the homogeneous truck set M, drone set N, truck volume Q1 and drone volume Q2 that stop at the warehouse;

[0016] The node set includes the warehouse node O and the customer point set V that the drone transfer station is responsible for. C , covering point set V L 、UAV Site Set V S , the set of customer points that coverage point i is responsible for, and the set of transportation tools

[0017] Node information includes customer i demand q i , the distance d between any two different points i and j in V ij , (i, j)∈A:={(i, j)|i, j∈V, i≠j}, truck unit distance transportation cost c1, drone unit distance transportation cost c2.

[0018] Preferably, the vehicle-machine collaborative delivery model includes an objective function, specifically as follows:

[0019]

[0020] in:

[0021]

[0022] Preferably, the vehicle-machine collaborative delivery model includes constraints, specifically as follows:

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] Constraints 1, 3, and 5 are used to constrain trucks. Each transfer station can only be visited once and by only one truck.

[0034] Constraints 2, 4, and 6 are used to constrain drones. Each drone client point can only be visited once and by only one drone.

[0035] Constraint 7 is used to eliminate the truck subloop;

[0036] Constraint 8 is used to eliminate the drone sub-loop;

[0037] Constraint 9 indicates that the total load of the truck cannot exceed its upper limit;

[0038] Constraint 10 means that the total payload of the drone cannot exceed its upper limit.

[0039] Preferably, the two-stage heuristic algorithm for solving the vehicle-machine collaborative delivery model includes:

[0040] Phase 1:

[0041] S311. Divide the belonging set and determine the needs of each transfer station;

[0042] S312. Based on the needs of each transfer station and under the premise of satisfying the truck capacity constraint, the first-level truck route is generated using the mileage saving method. The specific steps are as follows:

[0043] S312a, using the mileage saving algorithm to calculate the mileage saved between transfer stations, and sort and store them in the linked list CW<>, where the attributes include transfer station a1, transfer station a2, and mileage saved;

[0044] S312b, access the mileage saving table CW<> based on the idea of ​​sequential mileage saving algorithm, and proceed to step S312e;

[0045] S312c: Determine whether the cumulative route demand satisfies the vehicle volume constraint. If so, proceed to step S312d; if not, return to step S312b.

[0046] S312d, store the transfer station number in the corresponding truck route; then return to step S312b;

[0047] S312e: Determine whether CW<> has completed the visit. If not, proceed to step S312c. If so, terminate and output the first-level truck path.

[0048] S313. Shuffle the customer codes and randomly generate a second-level drone path;

[0049] The second-level UAV path is used as the initial solution, and the initial solution is optimized using the improved adaptive large neighborhood algorithm;

[0050] Phase 2:

[0051] S321, initialize the current solution, set the current number of iterations Nc = 0, the maximum number of iterations Nc max ;

[0052] S322, Nc≥Nc max Is it true? If so, output the current solution as the optimal solution. Otherwise, execute the next step. The optimal solution is the optimized drone path.

[0053] S323. Using the inverse of the total cost as the fitness, calculate the fitness of each individual in the current solution. Each time, select the top N individuals with the highest fitness from the population according to the elite selection strategy and perform the following operations:

[0054] S324. Selecting operator pairs based on roulette rules, wherein the operator pairs include nine operator pairs consisting of three destruction operators and three repair operators.

[0055] Destruction operators include:

[0056] ① Random removal of the destruction operator: randomly select a transfer station and randomly remove the customers provided by the transfer station;

[0057] ② Single-path removal destruction operator: randomly select a customer on a delivery path and remove them until q customers are removed;

[0058] ③ Association removal destruction operator: randomly generates a customer to be removed, calculates the association index between this customer point and other customer points. The association index considers the inverse of the distance d between the two customer points, and prioritizes the removal of customer nodes with low association.

[0059] Repair operators include:

[0060] ① Randomly insert repair operators: select any position on any path, and the solution after insertion satisfies the capacity constraint;

[0061] ② Greedy insertion repair operator: randomly select a path and insert all the points to be inserted at the position that increases the distance the least after insertion;

[0062] ③ Travel Insertion Repair Operator: It traverses each customer to be inserted, calculates the optimal insertion position and the corresponding minimum insertion cost, selects the customer with the lowest insertion cost and the insertion position among all customers, and performs the insertion operation;

[0063] S325. Based on the operator pair selected in the previous step, perform destruction and repair operations on the current solution to obtain a new solution. Determine whether the new solution is better than the current solution. If so, accept the new solution. Otherwise, accept the new solution with simulated annealing probability.

[0064] S326. Update the weight of the operator pair, Nc=Nc+1, and return to step S322.

[0065] Preferably, in step S325, before determining whether the new solution is better than the current solution, the method further includes determining whether to execute a rental strategy, wherein the rental strategy includes:

[0066] When there are idle drones at a drone transfer station, the drones at that transfer station are randomly leased to other transfer stations. If the total path cost after leasing is less than the total path cost before leasing, the leasing strategy is implemented; otherwise, the leasing strategy is implemented with simulated annealing probability. During the leasing process, each transfer station has at least one drone.

[0067] Preferably, in step S326, the updating of the weights of the operator pairs includes:

[0068]

[0069] Where: d，z is the operator weight, S d,z is the operator fraction, μ d,z is the number of times the operator is used and ρ is the weight update coefficient;

[0070] The operator score is calculated as follows:

[0071] At the beginning of the calculation, each operator has a score. During each calculation, the operator scores are as follows:

[0072] 1) The new solution updates the current local optimal solution and adds a point;

[0073] 2) The new solution updates the current global optimal solution and adds b points;

[0074] 3) If the new solution is worse than the current solution, but is accepted according to the algorithm's acceptance criteria, c points will be added;

[0075] 4) If the new solution is worse than the current solution and is not accepted, add d points;

[0076] Among them, a>b>c>d.

[0077] In a second aspect, the present invention provides a vehicle-cabinet collaborative delivery path optimization system, comprising:

[0078] Data acquisition module, used to obtain collaborative delivery task data with coverage points;

[0079] A model building module is used to build a vehicle-machine collaborative delivery model with a two-level path and coverage points based on collaborative delivery task data, with the goal of minimizing the total delivery cost;

[0080] The solution module is used to solve the vehicle-machine collaborative delivery model through a two-stage heuristic algorithm to obtain the optimized path.

[0081] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program for a vehicle-cabinet collaborative delivery path optimization system, wherein the computer program enables a computer to execute the vehicle-cabinet collaborative delivery path optimization method as described above.

[0082] In a fourth aspect, the present invention provides an electronic device, comprising:

[0083] one or more processors;

[0084] Memory; and

[0085] One or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include methods for executing the vehicle-cabinet collaborative delivery path optimization method as described above.

[0086] (3) Beneficial effects

[0087] The present invention provides a method and system for optimizing vehicle-cabinet collaborative delivery paths. Compared with existing technologies, it has the following advantages:

[0088] The present invention first obtains collaborative distribution task data with coverage points; then, based on the collaborative distribution task data, a vehicle-machine collaborative distribution model with a two-level path with coverage points is constructed with the goal of minimizing the total distribution cost; finally, the vehicle-machine collaborative distribution model is solved through a two-stage heuristic algorithm to obtain an optimized path. The vehicle-machine collaborative distribution model with a two-level path with coverage points constructed in the present invention can describe the vehicle-machine collaborative distribution problem in the two-level path problem while considering the introduction of the coverage point option, which is more in line with the actual process of vehicle-machine collaborative last-mile distribution. By solving the model, the optimized path obtained can effectively improve distribution efficiency and reduce terminal distribution costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0090] Figure 1 This is a block diagram of a vehicle-cabinet collaborative delivery path optimization method according to an embodiment of the present invention;

[0091] Figure 2 Flowchart of the improved adaptive large neighborhood algorithm in an embodiment of the present invention;

[0092] Figure 3 Schematic diagram of the operation process of randomly removing a destruction operator in an embodiment of the present invention;

[0093] Figure 4 Schematic diagram of the operation process of removing a destruction operator on a single path in an embodiment of the present invention;

[0094] Figure 5 Schematic diagram of the operation process of the correlation removal destruction operator in an embodiment of the present invention;

[0095] Figure 6 Schematic diagram of the operation process of randomly inserting a repair operator in an embodiment of the present invention;

[0096] Figure 7 Schematic diagram of the operation process of greedy insertion repair operator in an embodiment of the present invention;

[0097] Figure 8 Schematic diagram of the operation process of the travel insertion repair operator in an embodiment of the present invention;

[0098] Figure 9 Schematic diagram of the exchange operation process of the rental strategy in an embodiment of the present invention. DETAILED DESCRIPTION

[0099] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0100] The embodiments of the present application solve the technical problem of low delivery efficiency of existing methods by providing a vehicle-cabinet collaborative delivery path optimization method and system, improve the vehicle-cabinet collaborative delivery efficiency, and reduce terminal delivery costs.

[0101] The technical solution in the embodiments of the present application is to solve the above technical problems, and the overall idea is as follows:

[0102] Existing research on the two-echelon vehicle routing problem with drones has explored various perspectives and achieved considerable success. However, the following issues remain: First, existing solutions primarily focus on minimizing delivery costs without considering different customer service options (including home delivery and delivery to covered points); second, existing solutions for the two-echelon routing problem fail to consider the possibility of renting drones between stations; and third, existing solutions lack effective heuristic algorithms for the two-echelon vehicle-machine collaborative delivery problem with covered points.

[0103] Based on the above problems, the present invention proposes a research topic for a two-level vehicle routing problem with coverage points (such as express lockers or Cainiao stations). In the two-level vehicle routing problem with express lockers, goods are transferred from a central warehouse to transfer stations (drone stations and coverage points), and customers receive services through coverage points or drones. An integer programming model with minimizing delivery cost as the objective function is established, and an adaptive large neighborhood algorithm is designed for solving it, improving both solution and delivery efficiency.

[0104] The embodiment of the present invention provides a method for optimizing the collaborative delivery path of a vehicle cabinet. Figure 1 Shown, including:

[0105] S1. Obtain collaborative delivery task data with coverage points;

[0106] S2. Based on the collaborative delivery task data, a vehicle-machine collaborative delivery model with a two-level path and coverage points is constructed with the goal of minimizing the total delivery cost.

[0107] S3. Solve the vehicle-machine collaborative delivery model through a two-stage heuristic algorithm to obtain the optimized path.

[0108] The vehicle-machine collaborative delivery model with a two-level path and coverage points constructed in an embodiment of the present invention can describe the vehicle-machine collaborative delivery problem in the two-level path problem while considering the introduction of the coverage point option. It is more in line with the actual process of vehicle-machine collaboration in the last mile delivery. By solving the model, the optimized path obtained can effectively improve the delivery efficiency and reduce the terminal delivery cost.

[0109] The following is a detailed description of each step:

[0110] In step S1, collaborative delivery task data with coverage points is obtained. The specific implementation process is as follows:

[0111] In this embodiment of the present invention, obtaining collaborative delivery task data with coverage points mainly includes the following data:

[0112] The distribution network, i.e., the directed graph formed between warehouses, drone stations, coverage points, and customer locations, is represented by G = (V, A), where V represents the set of all nodes. A represents the set of all arcs, that is, the set of all delivery paths, A = {(i, j) | i, j∈V, i≠j}. It should be noted that the coverage points in the embodiment of the present invention take the express cabinet coverage points as an example.

[0113] The transportation tool set includes a set M of homogeneous trucks docked at the warehouse, a set N of drones, the truck volume Q1, and the drone volume Q2.

[0114] Node set and node information, where the node set includes warehouse node O and the customer point set V that the drone transfer station is responsible for C , Express cabinet coverage point set V L 、UAV Site Set V S , express locker i is responsible for the collection of customer points and transportation tools Node information includes customer i demand q i , the distance d between any two different points i and j in V ij , (i, j)∈A:={(i, j)|i, j∈V, i≠j}, express cabinet unit distance transportation cost l i , truck unit distance transportation cost c1, drone unit distance transportation cost c2.

[0115] In step S2, based on the collaborative delivery task data, a vehicle-machine collaborative delivery model with a two-level path and coverage points is constructed with the goal of minimizing the total delivery cost. The specific implementation process is as follows:

[0116] In the two-level vehicle routing problem with coverage points, trucks collaborate with drones and parcel lockers to complete the last-mile delivery task. Trucks depart from the warehouse, travel to drone stations or parcel locker coverage points, and transfer the cargo from the truck to a transfer station. Each customer is served by a drone or parcel locker that takes off from a drone station. Drones between drone stations can be rented, and the goal is to find a delivery route that minimizes total delivery cost.

[0117] Since the customer point's affiliation with the transfer station has been determined at the beginning of model planning, and the path between the self-pickup cabinet and the customer point belonging to it is uniquely determined, with no decision space, the self-pickup cost can be extracted from the total cost.

[0118] The self-collection costs are as follows:

[0119]

[0120] Based on the above collaborative delivery tasks, the following model including objective function and constraints is constructed:

[0121] The objective function of this model consists of two parts: the first part is the truck transportation cost, and the second part is the drone transportation cost.

[0122]

[0123] in:

[0124]

[0125] Constraints include:

[0126]

[0127]

[0128]

[0129]

[0130]

[0131]

[0132]

[0133]

[0134]

[0135]

[0136] Constraints 1, 3, and 5 are used to constrain trucks. Each transfer station can only be visited once and by only one truck.

[0137] Constraints 2, 4, and 6 are used to constrain drones. Each drone client point can only be visited once and by only one drone.

[0138] Constraint 7 is used to eliminate the truck subloop;

[0139] Constraint 8 is used to eliminate the drone sub-loop;

[0140] Constraint 9 indicates that the total load of the truck cannot exceed its upper limit;

[0141] Constraint 10 means that the total payload of the drone cannot exceed its upper limit.

[0142] In step S3, the vehicle-machine collaborative delivery model is solved through a two-stage heuristic algorithm to obtain the optimized path. The specific implementation process is as follows:

[0143] This embodiment of the present invention designs a two-stage solution algorithm. In the first stage, customers are divided into belonging sets based on the greedy principle, and the first-level truck routes are generated based on the mileage saving algorithm. The random second-level drone routes are used as the initial solution. In the second stage, the improved adaptive large neighborhood algorithm is used to optimize the initial solution to improve the solution quality. The details are as follows:

[0144] Phase 1:

[0145] S311. Divide the belonging set and determine the requirements of each transfer station:

[0146] S311a. Calculate the distance from each customer i to each transfer station to obtain a distance matrix.

[0147] S311b. Assign customers to the nearest transfer station, subject to the capacity constraints of the drone stations. Assign each customer to the nearest common distribution center whenever possible. If the nearest common distribution center is already full, assign the customer to the next closest common distribution center. Repeat this step until all customers are assigned.

[0148] S311c. Determine the demand of each transfer station based on the demand of the customers assigned to each transfer station.

[0149] S312: The first-level path is a single-depot VRP problem. Based on the needs of each transfer station and under the premise of meeting the truck capacity constraint, the first-level truck path is formed using the mileage saving method. The specific steps are as follows:

[0150] S312a. Calculate the saved mileage between transfer stations using the mileage saving algorithm, and sort and store them in a linked list CW<>, where the attributes include transfer station a1, transfer station a2, and saved mileage.

[0151] S312b, access the saved mileage table CW<> based on the idea of ​​the sequential saved mileage algorithm, and enter step S312e.

[0152] S312c: Determine whether the cumulative route demand meets the vehicle volume constraint. If so, proceed to step S312d; if not, return to step S312b.

[0153] S312d, store the transfer station number in the corresponding truck route; then return to step S312b.

[0154] S312e: Determine whether CW<> has completed the visit. If not, proceed to step S312c. If so, terminate and output the first-level truck path.

[0155] S313,The secondary path is a multi-factory MDVRP problem. The customer codes of the secondary path are shuffled and the second-level drone path is randomly generated.

[0156] The second-level UAV path is used as the initial solution, and the initial solution is optimized using the improved adaptive large neighborhood algorithm.

[0157] The second stage, such as Figure 2 As shown:

[0158] S321, initialize the current solution, set the current number of iterations Nc = 0, the maximum number of iterations Nc max In the embodiment of the present invention, Nc max =500;

[0159] S322, Nc≥Nc max Is it true? If so, output the current solution as the optimal solution. Otherwise, execute the next step. The optimal solution is the optimized drone path. The first-level truck path and the optimized drone path constitute the optimized path.

[0160] S323. Calculate the fitness of each individual in the current solution and express it as the inverse of the target total cost, i.e., fit = 1 / C. Each time, select the top N individuals with the highest fitness from the population according to the elite selection strategy and perform the following operations (it should be noted that, in this embodiment of the present invention, the top 1 / 3 individuals in the population are selected):

[0161] S324. Selecting operator pairs based on roulette rules, wherein the operator pairs include nine operator pairs consisting of three destruction operators and three repair operators.

[0162] During the algorithm iteration process, the current solution must first be "destroyed," meaning that q customers are removed from the solution. A typical approach is to calculate a special ranking metric for the customers currently placed in secondary facilities, sort these customers, and then remove them from the set R based on the ranking until q customers have been removed. The destruction operator includes:

[0163] ①The operation process of randomly removing the destruction operator is as follows: Figure 3 As shown in the figure, first a transfer station is randomly selected, and then the customers served by the transfer station are randomly selected for removal. Since this destruction operator does not use any information, it often produces a group of not-so-good customers to be removed. However, this random removal method helps to search for diverse solutions and escape from the local optimum.

[0164] ② The operation process of removing the destruction operator by a single path is as follows: Figure 4 As shown, randomly select a customer on a delivery path and remove them until q customers are removed.

[0165] ③The operation process of the correlation removal destruction operator is as follows Figure 5As shown in the figure, first a customer to be removed is randomly generated, and then the correlation index between this customer point and other customer points is calculated. The correlation index considers the inverse of the distance d between the two customer points, and customer nodes with low correlation are removed first.

[0166] The repair operator performs repair operations based on the set of removed customers and the destroyed solution. After the customer point is inserted, it is necessary to check whether the constraints are met. The following are the three repair operators used in the algorithm:

[0167] ①The operation process of randomly inserting the repair operator is as follows: Figure 6 As shown in Figure 2, any position on any path can be selected and inserted as long as the solution after insertion meets the capacity constraint.

[0168] ②The operation process of greedy insertion repair operator is as follows Figure 7 As shown, a path is randomly selected and all the points to be inserted are inserted at the position where the distance increased the least after insertion.

[0169] ③ The operation process of the travel insertion repair operator is as follows Figure 8 As shown, each customer to be inserted is traversed, the optimal insertion position and the corresponding minimum insertion cost are calculated, and then the customer with the minimum insertion cost and the insertion position among all customers are selected to perform the insertion operation.

[0170] A combination of three destruction operators and three repair operators is randomly selected based on roulette rules. The destruction operator and the repair operator can form 9 operator pairs.

[0171] S325. Based on the operator pair selected in the previous step, perform destruction and repair operations on the current solution to obtain a new solution f(x). Determine whether the new solution is better than the current solution. If so, accept the new solution; otherwise, accept the new solution with simulated annealing probability.

[0172] The Metropolis criterion is often expressed as:

[0173]

[0174] Among them, E n Indicates the energy of the previous state; E n+1 Indicates the energy of the current state.

[0175] In practical applications, the reasonable redistribution of drones at each transfer station is considered to improve delivery efficiency. Therefore, the concept of "rental" is proposed in the optimization process to exchange drones between paths. When there are idle drones at a drone transfer station, a random attempt is made to "rent" the drone of the transfer station to another transfer station. If the total path cost becomes lower after the lease, the rental strategy is implemented; if the cost becomes higher, the rental strategy is implemented with simulated annealing probability. During the rental process, it is ensured that each transfer station has at least one drone, with no upper limit. The exchange operation process of the rental strategy is as follows: Figure 9 shown.

[0176] S326. Update the weight of the operator pair, Nc=Nc+1, and return to step S322.

[0177] During each search, operator pairs are selected based on roulette rules to increase the diversity of operator pair selection. At the beginning of the calculation, each operator pair has the same weight and score, and the initial score is 1. Scoring is based on the performance of the operator pair, with higher scores indicating better performance. The operator pair scores during each calculation are as follows:

[0178] 1) If the new solution updates the current local optimal solution, 5 points will be added.

[0179] 2) The new solution updates the current global optimal solution and adds another 5 points.

[0180] 3) If the new solution is worse than the current solution but is still accepted according to the algorithm's acceptance criteria, 2 points will be added.

[0181] 4) If the new solution is worse than the current solution and is not accepted, add 0.8 points.

[0182] The weight of the current operator pair participating in the roulette selection is calculated based on the average score of each round. In addition, a weight update coefficient ρ is set to avoid the convergence speed falling into the local optimum too quickly, and the weight ω corresponding to the operator pair (d, z) is updated. d，z The specific method is as follows:

[0183]

[0184] Where: d，z is the operator weight, S d,z is the operator fraction, μ d,z is the number of times the operator is used.

[0185] An embodiment of the present invention further provides a vehicle-cabinet collaborative delivery path optimization system, comprising:

[0186] Data acquisition module, used to obtain collaborative delivery task data with coverage points;

[0187] A model building module is used to build a vehicle-machine collaborative delivery model with a two-level path and coverage points based on collaborative delivery task data, with the goal of minimizing the total delivery cost;

[0188] The solution module is used to solve the vehicle-machine collaborative delivery model through a two-stage heuristic algorithm to obtain the optimized path.

[0189] It can be understood that the vehicle cabinet collaborative distribution path optimization system provided by the embodiment of the present invention corresponds to the above-mentioned vehicle cabinet collaborative distribution path optimization method. The explanation, examples, beneficial effects and other parts of its relevant contents can refer to the corresponding contents in the vehicle cabinet collaborative distribution path optimization method, and will not be repeated here.

[0190] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program for optimizing the vehicle-cabinet collaborative delivery path, wherein the computer program enables a computer to execute the vehicle-cabinet collaborative delivery path optimization method as described above.

[0191] An embodiment of the present invention further provides an electronic device, including:

[0192] one or more processors;

[0193] Memory; and

[0194] One or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include methods for executing the vehicle-cabinet collaborative delivery path optimization method as described above.

[0195] In summary, compared with the existing technology, the present invention has the following beneficial effects:

[0196] 1. The vehicle-machine collaborative delivery model with a two-level path and coverage points constructed in an embodiment of the present invention can describe the vehicle-machine collaborative delivery problem in the two-level path problem while considering the introduction of the coverage point option. This model is more in line with the actual process of vehicle-machine collaborative delivery for the last mile. By solving this model, the optimized path obtained can effectively improve delivery efficiency and reduce terminal delivery costs.

[0197] 2. In the process of route optimization, the embodiment of the present invention takes into account the fact that drones can be rented between sites. The resulting optimized route is more in line with the actual situation, further improving delivery efficiency.

[0198] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0199] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A vehicle-cabinet collaborative delivery path optimization method, characterized in that: The method comprises: S1. Obtain collaborative delivery task data with coverage points; S2. Based on the collaborative delivery task data, a vehicle-machine collaborative delivery model with a two-level path and coverage points is constructed with the goal of minimizing the total delivery cost. S3. Solve the vehicle-machine collaborative delivery model through a two-stage heuristic algorithm to obtain the optimized path; The two-stage heuristic algorithm for solving the vehicle-machine collaborative delivery model includes: Phase 1: S311. Divide the belonging set and determine the needs of each transfer station; S312. Based on the needs of each transfer station and under the premise of satisfying the truck capacity constraint, the first-level truck route is generated using the mileage saving method. The specific steps are as follows: S312a, using the mileage saving algorithm to calculate the mileage saved between transfer stations, and sort and store them in the linked list CW<>, where the attributes include transfer station a1, transfer station a2, and mileage saved; S312b, access the mileage saving table CW<> based on the idea of ​​sequential mileage saving algorithm, and proceed to step S312e; S312c: Determine whether the cumulative route demand satisfies the vehicle volume constraint. If so, proceed to step S312d; if not, return to step S312b. S312d, store the transfer station number in the corresponding truck route; then return to step S312b; S312e, determine whether CW<> has been accessed. If not, proceed to step S312c; if so, terminate and output the first-level truck path. S313. Shuffle the customer codes and randomly generate a second-level drone path; The second-level UAV path is used as the initial solution, and the initial solution is optimized using the improved adaptive large neighborhood algorithm; Phase 2: S321, initialize the current solution, set the current number of iterations Nc = 0, the maximum number of iterations Nc max ; S322, Nc≥Nc max Is it true? If so, output the current solution as the optimal solution. Otherwise, execute the next step. The optimal solution is the optimized drone path. S323. Calculate the fitness of each individual in the current solution using the inverse of the total cost as the fitness. Each time, select the top N individuals with the highest fitness from the population according to the elite selection strategy and perform the following operations: S324, selecting an operator pair based on roulette rules; S325. Based on the operator pair selected in the previous step, perform destruction and repair operations on the current solution to obtain a new solution. Determine whether the new solution is better than the current solution. If so, accept the new solution. Otherwise, accept the new solution with simulated annealing probability. S326. Update the weight of the operator pair, Nc=Nc+1, and return to step S322.

2. The vehicle-cabinet collaborative delivery path optimization method according to claim 1, characterized in that: The collaborative delivery task data with coverage points includes: a delivery network, a transport tool set, a node set, and node information; The distribution network refers to a directed graph formed between warehouses, drone sites, coverage points, and customer locations, represented by G = (V, A), where V represents the set of all nodes. A represents the set of all arcs, that is, the set of all delivery paths, A = {(i, j) | i, j∈V, i≠j}; The transportation tool set includes the homogeneous truck set M, drone set N, truck volume Q1 and drone volume Q2 that stop at the warehouse; The node set includes the warehouse node O and the customer point set V that the drone transfer station is responsible for. C , covering point set V L 、UAV Site Set V S , the set of customer points that coverage point i is responsible for, and the set of transportation tools Node information includes customer i demand q i , the distance d between any two different points i and j in V ij ,(i,j)∈A:={(i,j)|i,j∈V,i≠j}, truck unit distance transportation cost c1, drone unit distance transportation cost c2.

3. The vehicle-cabinet collaborative delivery path optimization method according to claim 2, characterized in that: The vehicle-machine collaborative delivery model includes an objective function, which is as follows: in:

4. The vehicle-cabinet collaborative delivery path optimization method according to claim 3, characterized in that: The vehicle-machine collaborative delivery model includes constraints, which are as follows: ∑ i∈O∪VL∪Vs ∑ k∈M x ijk =1,j∈V L ∪V S (1) ∑ i∈VC∪Vs ∑ k∈N x ijk =1,j∈V C (2)∑ i∈O∪VL∪Vs x ijk =∑ i∈O∪VL∪Vs x jik ,j∈O∪V L ∪V S ,k∈M(3) ∑ i∈Vc∪Va x ijk =∑ i∈VC∪Vs x jik ,j∈V C ∪V S ,k∈N (4) ∑ i∈VL∪Vs ∑ j∈O x ijk ≤1,k∈M (5) ∑ i∈VC ∑ j∈Vs x ijk ≤1,k∈N (6)∑ k∈M u ik -∑ k∈M u ik +(L+S)∑ k∈M x ijk ≤L+S-1,i,j∈V L ∪V S (7)∑ k∈N u ik -∑ k∈N u ik +C∑ k∈N x ijk ≤C-1,i,j∈V C (8) ∑ i∈O∪VL∪Vs ∑ j∈VL∪Vs x ijk q j ≤Q1,k∈M (9) ∑ i∈VC∪Vs ∑ j∈Vc x ijk q j ≤Q2,k∈N(10) Constraints 1, 3, and 5 are used to constrain trucks. Each transfer station can only be visited once and by only one truck. Constraints 2, 4, and 6 are used to constrain drones. Each drone client point can only be visited once and by only one drone. Constraint 7 is used to eliminate the truck subloop; Constraint 8 is used to eliminate the drone sub-loop; Constraint 9 indicates that the total load of the truck cannot exceed its upper limit; Constraint 10 means that the total payload of the drone cannot exceed its upper limit.

5. The vehicle-cabinet collaborative delivery path optimization method according to any one of claims 1 to 4, characterized in that: The operator pairs include 9 operator pairs consisting of 3 destruction operators and 3 repair operators; Destruction operators include: ① Random removal of the destruction operator: randomly select a transfer station and randomly remove the customers provided by the transfer station; ② Single-path removal destruction operator: randomly select a customer on a delivery path and remove them until q customers are removed; ③ Association removal destruction operator: randomly generates a customer to be removed, calculates the association index between this customer point and other customer points. The association index considers the inverse of the distance d between the two customer points, and prioritizes the removal of customer nodes with low association. Repair operators include: ① Randomly insert repair operators: select any position on any path, and the solution after insertion satisfies the capacity constraint; ② Greedy insertion repair operator: randomly select a path and insert all the points to be inserted at the position that increases the distance the least after insertion; ③ Travel Insertion Repair Operator: It traverses each customer to be inserted, calculates the optimal insertion position and the corresponding minimum insertion cost, selects the customer with the lowest insertion cost and insertion position among all customers, and performs the insertion operation.

6. The vehicle-cabinet collaborative delivery path optimization method according to claim 5, characterized in that: In step S325, before determining whether the new solution is better than the current solution, the method further includes determining whether to execute a rental strategy, wherein the rental strategy includes: When there are idle drones at a drone transfer station, the drones at that transfer station are randomly leased to other transfer stations. If the total path cost after leasing is less than the total path cost before leasing, the leasing strategy is implemented; otherwise, the leasing strategy is implemented with simulated annealing probability. During the leasing process, each transfer station has at least one drone.

7. The vehicle-cabinet collaborative delivery path optimization method according to claim 5, characterized in that: In step S326, the weight of the updating operator pair includes: Where: d，z is the operator weight, S d,z is the operator fraction, μ d,z is the number of times the operator is used and ρ is the weight update coefficient; The operator score is calculated as follows: At the beginning of the calculation, each operator pair has the same score. During each calculation, the operator pairs are scored as follows: 1) The new solution updates the current local optimal solution and adds a point; 2) The new solution updates the current global optimal solution and adds b points; 3) If the new solution is worse than the current solution, but is accepted according to the algorithm's acceptance criteria, c points will be added; 4) If the new solution is worse than the current solution and is not accepted, add d points; Among them, a>b>c>d.

8. A vehicle-cabinet collaborative delivery path optimization system, characterized in that: include: Data acquisition module, used to obtain collaborative delivery task data with coverage points; A model building module is used to build a vehicle-machine collaborative delivery model with a two-level path and coverage points based on collaborative delivery task data, with the goal of minimizing the total delivery cost; The solution module is used to solve the vehicle-machine collaborative delivery model through a two-stage heuristic algorithm to obtain the optimized path; The two-stage heuristic algorithm for solving the vehicle-machine collaborative delivery model includes: Phase 1: S311. Divide the belonging set and determine the needs of each transfer station; S312. Based on the needs of each transfer station and under the premise of satisfying the truck capacity constraint, the first-level truck route is generated using the mileage saving method. The specific steps are as follows: S312a, using the mileage saving algorithm to calculate the mileage saved between transfer stations, and sort and store them in the linked list CW<>, where the attributes include transfer station a1, transfer station a2, and mileage saved; S312b, access the mileage saving table CW<> based on the idea of ​​sequential mileage saving algorithm, and proceed to step S312e; S312c: Determine whether the cumulative route demand satisfies the vehicle volume constraint. If so, proceed to step S312d; if not, return to step S312b. S312d, store the transfer station number in the corresponding truck route; then return to step S312b; S312e: Determine whether CW<> has completed the visit. If not, proceed to step S312c. If so, terminate and output the first-level truck path. S313. Shuffle the customer codes and randomly generate a second-level drone path; The second-level UAV path is used as the initial solution, and the initial solution is optimized using the improved adaptive large neighborhood algorithm; Phase 2: S321, initialize the current solution, set the current number of iterations Nc = 0, the maximum number of iterations Nc max ; S322, Nc≥Nc max Is it true? If so, output the current solution as the optimal solution. Otherwise, execute the next step. The optimal solution is the optimized drone path. S323. Using the inverse of the total cost as the fitness, calculate the fitness of each individual in the current solution. Each time, select the top N individuals with the highest fitness from the population according to the elite selection strategy and perform the following operations: S324, selecting an operator pair based on roulette rules; S325. Based on the operator pair selected in the previous step, perform destruction and repair operations on the current solution to obtain a new solution. Determine whether the new solution is better than the current solution. If so, accept the new solution. Otherwise, accept the new solution with simulated annealing probability. S326. Update the weight of the operator pair, Nc=Nc+1, and return to step S322.

9. A computer-readable storage medium, characterized in that It stores a computer program for a vehicle-cabinet collaborative delivery path optimization system, wherein the computer program enables a computer to execute the vehicle-cabinet collaborative delivery path optimization method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: include: one or more processors; Memory; as well as One or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the programs include programs for executing the vehicle-cabinet collaborative distribution path optimization method as described in any one of claims 1 to 7.

Images