Site selection-path decision method for joint delivery of truck and unmanned vehicle and related device

Abstract

The invention provides a site selection-path decision method for joint delivery of a truck and an unmanned vehicle and a related device, and relates to the technical field of delivery planning. The method comprises the following steps: when a distribution task is received, obtaining unmanned vehicle distribution candidate distribution points and service ranges thereof according to the distribution task; acquiring customer requirements according to the distribution task, and dynamically adjusting a distribution point selection threshold to judge whether each candidate distribution point is started or not; for the started gathering and distribution points, through an improved adaptive large neighborhood search algorithm, a truck and unmanned vehicle joint distribution path is optimized, and the total distribution cost is minimized; and calculating and outputting the total distribution cost, the optimal path and the distribution point selection state. Aiming at a delivery mode of joint delivery of trucks and unmanned vehicles, an unmanned vehicle distribution point dynamic site selection strategy based on delivery requirements is provided, delivery tasks are flexibly distributed, the labor intensity of workers is reduced, and the delivery efficiency is improved.

Description

Technical Field

[0001] The present application relates to the technical field of distribution planning, and in particular to a site selection-path decision method and related devices for joint distribution by trucks and unmanned vehicles. Background Art

[0002] In recent years, with the rapid development of e-commerce, the demand for logistics and distribution has shown explosive growth. In the entire logistics and distribution process, terminal delivery is considered to be the most difficult link with the lowest consumer satisfaction. The choice of its strategy directly affects logistics efficiency, logistics costs and customer satisfaction. The traditional terminal delivery model relies on manual labor to perform a large number of repetitive delivery tasks, which not only leads to high workload for practitioners, but also causes high operating costs due to huge manpower investment. In addition, the problem of population aging is becoming increasingly serious, the supply of working-age labor is decreasing, and labor costs are continuously rising. The traditional distribution model has become increasingly difficult to meet the growing demand for terminal delivery.

[0003] In this context, intelligent delivery technologies represented by drones and unmanned vehicles provide new solutions for terminal logistics innovation. Unmanned delivery systems reduce the dependence of terminal delivery on manual labor through automation and intelligent equipment, have the ability to operate continuously around the clock, significantly improve delivery efficiency, reduce labor costs, and show broad application prospects. Current research hotspots are mainly focused on the collaborative delivery model of trucks and drones, but this solution still has limitations in high-density scenarios such as urban communities. For example, drones are limited by their load capacity and need to make multiple round trips in demand-intensive areas, resulting in longer flight distances and increased energy consumption and operating costs. At the same time, extreme weather can also affect their stability and applicability. In contrast, unmanned vehicles have obvious advantages in load capacity, cruising range, and environmental adaptability, and are more suitable for large-scale delivery in demand-intensive areas.

[0004] Therefore, how to flexibly allocate delivery tasks, reduce manual workload, and improve delivery efficiency has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The core of the present invention is to provide a site selection-route decision method and related devices for joint delivery of trucks and unmanned vehicles, so as to solve the problems of flexibly allocating delivery tasks, reducing manual workload and improving delivery efficiency.

[0006] In the first aspect, the present application provides a method for site selection and route decision-making for joint delivery of trucks and unmanned vehicles, which adopts the following technical solutions: A location selection and routing decision method for joint delivery of trucks and unmanned vehicles, comprising: When receiving a delivery task, obtaining candidate collection and distribution points for unmanned vehicle delivery and their service ranges according to the delivery task; Obtain customer requirements according to the delivery task, and dynamically adjust the selection threshold of the distribution point to determine whether each candidate distribution point is enabled; For the enabled distribution points, optimize the combined delivery path of trucks and autonomous vehicles through an improved adaptive large neighborhood search algorithm to minimize the total delivery cost; Calculate and output the total delivery cost, the optimal path, and the selection status of the distribution point. Among them, the total delivery cost includes the truck driving cost, the autonomous vehicle usage time cost, the driver labor cost, and the fixed usage cost of the distribution point.

[0007] Optionally, the dynamic adjustment of the distribution point selection threshold specifically includes: Calculate the difference between the autonomous vehicle delivery cost when the distribution point is enabled and the truck delivery cost when it is not enabled; Dynamically update the selection threshold according to the difference. If the autonomous vehicle delivery cost is lower, increase the selection threshold to increase the probability of enabling the distribution point, and vice versa, decrease the selection threshold; Judge whether the distribution point is enabled through the normal distribution function, and generate a distribution point selection status coding sequence.

[0008] Optionally, the improved adaptive large neighborhood search algorithm includes: Use a destruction operator to locally destroy the truck path, including random destruction, worst-cost destruction, and related destruction; Use a repair operator to reconstruct the destroyed path, including random repair, greedy repair, and regret repair; Design a neighborhood search operator for the autonomous vehicle path, including Swap, Insert, 2-opt, and or-opt operators; Introduce a simulated annealing acceptance criterion to probabilistically accept a worse solution and avoid the algorithm falling into a local optimum.

[0009] Optionally, the similarity calculation of the destruction operator is: Among them, and are weight coefficients, is the customer point to the customer point distance, and are the demand quantities of customer point and customer point respectively. If customer point and customer point are on the same path is 0, otherwise it is 1. Remove them from largest to smallest similarity until the specified destruction quantity is reached.

[0010] Optionally, the step of reconstructing the damaged path using the repair operator includes: Obtain the operator weights and select one of the three repair operators according to the operator weights by roulette method to repair the truck path to obtain a solution ; Among them, random repair includes: randomly selecting a customer point from the set of damaged customer points and inserting it into a random position in the truck delivery path until there are no remaining customer points to insert; Greedy repair includes: selecting a customer point from the set of damaged customer points, calculating the increment of the delivery cost when the customer point is inserted into different positions of different truck delivery paths, and selecting the insertion position with the smallest increment of the delivery cost and inserting it into that position until there are no remaining customer points to insert.

[0011] Regret repair includes: calculating the delivery cost at the optimal insertion position and the delivery cost at the sub-optimal insertion position for each customer in the set of damaged customer points, and taking the difference between the costs of the two positions as the regret value, and selecting the customer point with the largest regret value and inserting it into the optimal position.

[0012] Optionally, the step of designing the neighborhood search operator for the unmanned vehicle path includes: Select one of the four neighborhood search operators by roulette method to obtain the delivery path of the unmanned vehicle ; Among them, the Swap operator exchanges the positions of two random customer points in the unmanned vehicle delivery path; The Insert operator randomly selects two customer points in the unmanned vehicle delivery path , and inserts the customer point after inserting the customer point ; The 2-opt operator randomly selects 2 customer nodes , and reverses the customer positions between the customer node and the customer node ; The or-opt operator randomly selects two customer points in the unmanned vehicle delivery path , and inserts the customer point and the customer point after flipping them together into the customer point ;

[0013] Optionally, the mathematical model of the total delivery cost is: Among them, is the truck driving distance cost, is the unmanned vehicle usage time cost, is the driver labor cost, It is the fixed usage cost of the distribution point and satisfies the path flow conservation and capacity constraints.

[0014] In a second aspect, the present application provides a location - route decision - making device for combined distribution of trucks and autonomous vehicles, which executes the method described above and includes: A data acquisition module, configured to obtain candidate distribution points for autonomous vehicle distribution and their service scopes according to the distribution task when receiving the distribution task; A judgment module, configured to obtain customer demands according to the distribution task, dynamically adjust the threshold for selecting distribution points to judge whether each candidate distribution point is enabled; An algorithm module, for the enabled distribution points, optimizes the combined distribution path of trucks and autonomous vehicles through an improved adaptive large - neighborhood search algorithm to minimize the total distribution cost; A calculation module, configured to calculate and output the total distribution cost, the optimal path, and the selection status of the distribution points, where the total distribution cost includes the truck driving cost, the autonomous vehicle usage time cost, the driver labor cost, and the fixed usage cost of the distribution point.

[0015] In a third aspect, the present application provides a computer device, which includes: a memory and a processor. When the processor runs the computer instructions stored in the memory, it executes the method described above.

[0016] In a fourth aspect, the present application provides a computer - readable storage medium, including instructions. When the instructions run on a computer, the computer is made to execute the method described above.

[0017] In summary, the present application includes the following beneficial technical effects: When receiving a distribution task, the present application obtains candidate distribution points for autonomous vehicle distribution and their service scopes according to the distribution task; obtains customer demands according to the distribution task, dynamically adjusts the threshold for selecting distribution points to judge whether each candidate distribution point is enabled; for the enabled distribution points, optimizes the combined distribution path of trucks and autonomous vehicles through an improved adaptive large - neighborhood search algorithm to minimize the total distribution cost; calculates and outputs the total distribution cost, the optimal path, and the selection status of the distribution points. A dynamic location - selection strategy for autonomous vehicle distribution points based on distribution demands is proposed for the distribution mode of combined distribution of trucks and autonomous vehicles, which flexibly allocates distribution tasks, reduces the manual work intensity, and improves the distribution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic structural diagram of a computer device for the hardware operating environment involved in the solution of the embodiment of the present application.

[0019] Figure 2 It is a schematic flowchart of the first embodiment of the location - route decision - making method for combined distribution of trucks and autonomous vehicles of the present application.

[0020] Figure 3 It is a conceptual diagram of the site - route planning process for the joint distribution of trucks and driverless vehicles in this application.

[0021] Figure 4 It is a conceptual diagram of the site - selection strategy in this application.

[0022] Figure 5 It is a flow chart of the improved adaptive large - neighborhood search algorithm in this application.

[0023] Figure 6 It is a schematic diagram of the neighborhood operator. Figure 6 (a) is a schematic diagram of swap. Figure 6 (b) is a schematic diagram of Insert. Figure 6 (c) is a schematic diagram of 2 - opt. Figure 6 (d) is a schematic diagram of or - opt.

[0024] Figure 7 It is the site - route planning diagram of driverless vehicles for test cases C101, C201, R101, and RC202. Figure 7 (a) is the site - route planning diagram of C101. Figure 7 (b) is the site - route planning diagram of C201. Figure 7 (c) is the site - route planning diagram of R101. Figure 7 (d) is the site - route planning diagram of RC101.

[0025] Figure 8 It is a structural block diagram of the first embodiment of the site - route decision - making device for the joint distribution of trucks and driverless vehicles in this application. Detailed implementation manners

[0026] In order to make the objectives, technical solutions, and advantages of this application clearer, the following further details this application through the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not used to limit this application.

[0027] Refer to Figure 1 , Figure 1 It is a structural schematic diagram of the computer device for the hardware operating environment involved in the solution of the embodiment of this application.

[0028] As Figure 1As shown in the figure, a computer device may include: a processor 1001, such as a Central Processing Unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard). Optionally, the user interface 1003 may further include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed Random Access Memory (RAM), or a stable Non-Volatile Memory (NVM), such as a disk memory. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0029] Those skilled in the art can understand that Figure 1 the structure shown in does not constitute a limitation on the computer device, and it may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements.

[0030] As Figure 1 shown, in the memory 1005 as a storage medium, there may be included an operating system, a network communication module, a user interface module, and a location-path decision-making program for the joint distribution of trucks and unmanned vehicles.

[0031] In Figure 1 the computer device shown, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; in the present application, the processor 1001 and the memory 1005 may be arranged in the computer device. The computer device calls the location-path decision-making program stored in the memory 1005 through the processor 1001, and executes the location-path decision-making method for the joint distribution of trucks and unmanned vehicles provided in the embodiments of the present application.

[0032] The embodiments of the present application provide a location-path decision-making method for the joint distribution of trucks and unmanned vehicles. Referring to Figure 2 , Figure 2 is a schematic flowchart of the first embodiment of the location-path decision-making method for the joint distribution of trucks and unmanned vehicles in the present application.

[0033] In this embodiment, the location-path decision-making method for the joint distribution of trucks and unmanned vehicles includes the following steps: Step S10: When receiving a delivery task, obtain the candidate distribution collection and distribution points of the driverless vehicle and their service areas according to the delivery task.

[0034] In specific implementation, the term explanations in this embodiment include: VRP: An abbreviation for Vehicle Routing Problem, a classic optimization problem in operations research, aiming to design the optimal route for a fleet of vehicles to serve a group of customers at the lowest cost (such as distance, time) under the satisfaction of constraints (such as vehicle capacity, time window).

[0035] The Heuristic Algorithm is an algorithm that quickly finds a feasible solution to a problem through experience or intuitive rules. Its core goal is to find a "good enough" solution within a reasonable time, rather than guaranteeing a mathematically optimal solution. It is particularly suitable for complex optimization problems (such as NP-hard problems), where traditional exact algorithms (such as exhaustive search, dynamic programming) are difficult to directly apply due to excessive computational complexity.

[0036] Collection and distribution point: A site for trucks to unload goods, driverless vehicles to dock, and recharge driverless vehicles, similar to a transfer station.

[0037] It should be noted that the technical problems in the existing scenarios in this embodiment include: (1) In existing research, most adopt single-vehicle full-process delivery or the cooperation of trucks and drones. The former uses manual driving to complete the entire process of operation, facing great delivery pressure in areas with intensive demand. The latter, through the division of labor and cooperation between trucks and drones, although alleviates the manual delivery pressure to a certain extent, is limited by the load capacity of drones and needs to make frequent round trips to execute delivery tasks in areas with high-density demand, resulting in an increase in the charging frequency of equipment and ultimately reducing the overall delivery efficiency.

[0038] (2) Since driverless vehicles are relatively large in size and need to rely on collection and distribution points for storage and charging, in the existing scenario, for the transportation problem involving collection and distribution points, it is usually assumed that trucks only undertake the trunk transportation between collection and distribution points, and the customer delivery tasks are completely responsible for by the short-distance vehicles equipped at the collection and distribution points. When allocating tasks in this delivery mode, the along-the-way delivery capacity of trucks is not fully considered. When the customer point is in the middle area between two collection and distribution points, if only relying on the vehicles at the collection and distribution points to execute the delivery, the vehicles will inevitably make long-distance round trips, not only increasing the delivery distance but also limiting the space for path optimization.

[0039] (3) The current research generally adopts the fixed location selection mode for collection and distribution points, that is, after determining the location of the collection and distribution points, it maintains the operation state for a long time. There is a lack of research on the dynamic location selection of collection and distribution points for the dynamic changes in customer needs. In the actual logistics scenario, the customer distribution is complex and changeable, and the customer distribution has an important impact on the distribution efficiency and cost. Simply relying on the operation mode of fixed collection and distribution points is difficult to adapt to the complex and changeable customer distribution characteristics, resulting in insufficient operation flexibility.

[0040] Therefore, the problems to be solved in this embodiment are as follows: there are a large number of end - distribution tasks, and a large number of repetitive round - trips are required within a certain range. For example, there are a large number of residents in urban communities, and the daily distribution tasks are numerous. Relying solely on manual labor leads to a large distribution pressure, while drones are limited by their load - carrying capacity and cannot give full play to their advantages in such demand - intensive areas. At the same time, in the existing distribution mode with collection and distribution points, when facing the complex and changeable customer distribution, the distribution task allocation is not flexible enough, resulting in limited path optimization space.

[0041] In the specific implementation, to solve the problems of a large number of manual repetitive operations and low distribution efficiency in end - distribution, this embodiment adopts a combined distribution mode of trucks and unmanned vehicles to serve customers. The logistics system consists of a distribution center, candidate collection and distribution points for unmanned vehicle distribution, customers, trucks, and unmanned vehicles. The coordinates, demand quantities, required service times, truck capacities, unmanned vehicle capacities, coordinates and service scopes of candidate collection and distribution points for unmanned vehicle distribution of each customer are known. Multiple collection and distribution points for unmanned vehicle distribution are preset according to factors such as customer distribution and demand quantity. There are multiple unmanned vehicles at each collection and distribution point for distribution. The unmanned vehicle can only distribute customers within the service scope of the affiliated collection and distribution point, and customers outside the service scope of the collection and distribution point are distributed by trucks. During logistics distribution, first, according to the dynamic location selection strategy of unmanned vehicle collection and distribution points, a suitable collection and distribution point is selected as the collection and distribution point for unmanned vehicle distribution. On this basis, the truck loads goods and departs from the distribution center to distribute customers. After the truck arrives at the collection and distribution point for unmanned vehicle distribution, it unloads the goods of all customers within the service scope of this collection and distribution point, and then the unmanned vehicle distributes the customers in this area. If the candidate collection and distribution point for unmanned vehicle distribution is not selected, all customers within the distribution scope of this collection and distribution point are distributed by trucks. After the truck completes the distribution task, it returns to the distribution center, and after the unmanned vehicle completes the distribution task, it returns to the original departure collection and distribution point. The specific process is as Figure 3 shown.

[0042] It should be noted that the specific applicable scope in this embodiment corresponds to the following conditions: 1. The trucks are of the same type, and the unmanned vehicles are of the same type; 2. The demand quantity of each customer within the area of the collection and distribution point for unmanned vehicle distribution is less than the capacity of the unmanned vehicle; 3. Each customer can be served only once by one unmanned vehicle or one truck; 4. Each unmanned vehicle is enabled for distribution at most once, and each truck is enabled for distribution at most once.

[0043] In the specific implementation, the total distribution cost includes the truck driving cost, the truck driver’s salary cost, the truck’s time cost for serving customers and the truck’s time cost for serving distribution points, the unmanned vehicle’s time cost, and the fixed usage cost of distribution points.

[0044] Step S20: Obtain customer demand according to the delivery task, and dynamically adjust the distribution point selection threshold to determine whether each candidate distribution point is enabled.

[0045] It should be noted that the dynamic adjustment of the distribution point selection threshold specifically includes: calculating the difference between the delivery cost of the unmanned vehicle when the distribution point is enabled and the delivery cost of the truck when it is not enabled; dynamically updating the selection threshold according to the difference, if the unmanned vehicle delivery cost is lower, then raising the selection threshold to increase the probability of enabling the distribution point, otherwise lowering the selection threshold; judging whether the distribution point is enabled by the normal distribution function, and generating a distribution point selection status coding sequence.

[0046] In specific implementation, the dynamic location selection strategy for unmanned vehicle distribution points includes: The site selection of unmanned vehicle distribution points is a dynamic and cyclical optimization process. In the initial state, the selection threshold is set to 0.5, and the goal is to deliver to all customers in the enabled distribution point area at the lowest cost.

[0047] The specific steps of site selection strategy are as follows: Figure 4 shown.

[0048] exist Figure 4 The corresponding site selection process includes: Step 1: Calculate the delivery cost of the distribution point area. In the current iteration, the unmanned vehicle path in the distribution point area is optimized, and the cost is calculated in the following two ways based on the optimization results: ① Calculate the cost of the unmanned vehicle from the distribution point The cost of returning to the original distribution point after completing the delivery task for all customer points in the area, the fixed cost of using the distribution point, and the cost of trucks entering and leaving the distribution point from the customer point closest to the distribution point area. ② Calculate the sum of the costs of truck direct distribution to the distribution point (the distribution cost when the corresponding distribution point is selected); ② Calculate the distribution cost of truck direct distribution to the distribution point The sum of the distribution costs of all customer points within the distribution point area and the customer points closest to the distribution point area when driving in and out (the distribution cost when the corresponding distribution point is not selected).

[0049] Step 2: Update the selection threshold. By comparing the two costs in step 1, dynamically adjust the collection and distribution points according to the following formula: The selection threshold.

[0050] in, is the selection threshold of the collection and distribution points in this iteration ; is the selection threshold of the collection and distribution points in the previous iteration ; is the distribution cost of the truck at the collection and distribution point ; is the distribution cost of the driverless vehicle at the collection and distribution point ; is the difference in distribution costs between the truck and the driverless vehicle at the collection and distribution point ; is the attenuation coefficient

[0051] Step 3: Determine the selection status of the collection and distribution points. According to the selection threshold of the collection and distribution point in the current iteration, judge whether the collection and distribution point will be selected in the next iteration. If the threshold condition is met (that is, the function value that satisfies the normal distribution between [0,1] is less than the selection threshold ), then encode the collection and distribution point as 1 (indicating that the collection and distribution point is selected), otherwise, encode it as 0 (indicating that the collection and distribution point is not selected).

[0052] During the entire site selection optimization process, if the cost of using a driverless vehicle for distribution in a certain collection and distribution point area is lower than that of a truck after calculation, the selection threshold of this collection and distribution point will be increased, thereby increasing the probability of this collection and distribution point being selected in the next iteration. On the contrary, if the truck distribution cost is lower after calculation, the selection threshold of the collection and distribution point will be decreased, thereby reducing the probability of this collection and distribution point being selected. As the optimization process progresses, the selection status of the collection and distribution points gradually converges. For the collection and distribution points with lower driverless vehicle distribution costs when selected, they will be preferentially selected in subsequent iterations.

[0053] Step S30: For the enabled collection and distribution points, optimize the combined distribution path of the truck and the driverless vehicle through an improved adaptive large neighborhood search algorithm to minimize the total distribution cost.

[0054] In specific implementation, the improved adaptive large neighborhood search algorithm includes: Using a destruction operator to locally destroy the truck path, including random destruction, worst-cost destruction, and related destruction; Using a repair operator to reconstruct the destroyed path, including random repair, greedy repair, and regret repair; Designing a neighborhood search operator for the driverless vehicle path, including Swap, Insert, 2-opt, and or-opt operators; Introducing a simulated annealing acceptance criterion to accept worse solutions in a probabilistic manner to avoid the algorithm falling into a local optimum.

[0055] It should be noted that the similarity calculation of the destruction operator is as follows: Among them, and are weight coefficients, is the customer point to the customer point distance, and are respectively the demand quantities of the customer point and the customer point If the customer point and the customer point are on the same path is 0, otherwise it is 1, and they are removed from largest to smallest similarity until the specified destruction quantity is reached.

[0056] It can be understood that the step of reconstructing the damaged path by using the repair operator includes: Obtain the operator weight and select one of the three repair operators according to the operator weight by the roulette method to repair the truck path to obtain a solution ; Among them, random repair includes: randomly selecting a customer point from the set of damaged customer points and inserting it into a random position of the truck delivery path until there are no remaining insertable customer points; Greedy repair includes: selecting a customer point from the set of damaged customer points, calculating the increment of the delivery cost when the customer point is inserted into different positions of different truck delivery paths, and selecting the insertion position with the smallest increment of the delivery cost and inserting it into that position until there are no remaining insertable customer points.

[0057] Regret repair includes: calculating the delivery cost at the optimal insertion position and the delivery cost at the sub-optimal insertion position for each customer in the set of damaged customer points, and taking the difference between the costs of the two positions as the regret value, and selecting the customer point with the largest regret value and inserting it into the optimal position.

[0058] It should be noted that the step of designing the neighborhood search operator for the unmanned vehicle path includes: Select one of the four neighborhood search operators by the roulette method to obtain the delivery path of the unmanned vehicle ; Among them, the Swap operator exchanges the positions of two random customer points in the unmanned vehicle delivery path; The Insert operator randomly selects two customer points in the unmanned vehicle delivery path , and inserts the customer point after the customer point ; The 2-opt operator randomly selects 2 customer nodes , and the customer node The customer locations between it and the customer nodes are reversed; The or-opt operator randomly selects two customer points on the delivery route of the driverless vehicle , and the customer point and the customer point are flipped and then inserted together after the customer point

[0059] Step S40: Calculate and output the total delivery cost, the optimal route, and the selection status of the collection and distribution points. Among them, the total delivery cost includes the truck driving cost, the driverless vehicle usage time cost, the driver labor cost, and the fixed usage cost of the collection and distribution points.

[0060] In a specific implementation, the mathematical model of the total delivery cost is: Among them, is the truck driving distance cost, is the driverless vehicle usage time cost, is the driver labor cost, is the fixed usage cost of the collection and distribution points, and it satisfies the path flow conservation and capacity constraints.

[0061] In a specific implementation, the mathematical model of this embodiment is: (1) Sets. is the set of candidate collection and distribution points for driverless vehicle delivery; is the candidate collection and distribution point for driverless vehicle delivery and is the set of all customers within the service range; is the set of customers within the areas of all candidate collection and distribution points for driverless vehicle delivery, ; is the set of all customers outside the areas of candidate collection and distribution points for driverless vehicle delivery; is the set of all customers, ; is the set of all nodes within the logistics system, , where is the distribution center; is the set of trucks; is the set of driverless vehicles; is the set of all vehicles, .

[0062] (2) Symbols. is the truck capacity; is the driverless vehicle capacity; is the demand of customer ; is the sum of the demands of all customers within the service range of the driverless vehicle delivery collection and distribution point ; The distance from node to node . is the time when the vehicle arrives at node . is the time when the vehicle leaves from node . is the service time required by the customer . is the service time required by the distribution center . is the remaining drivable distance of the driverless vehicle when it arrives at the customer point . is the remaining drivable distance of the driverless vehicle when it leaves node . is the real-time load of the vehicle when it arrives at node . is the real-time load of the vehicle when it leaves node . is the driving speed of the truck is the driving speed of the driverless vehicle is the driving time of the truck from node to node . is the driving time of the driverless vehicle from node to node .

[0063] (3) Decision variables is a 0-1 variable. If the vehicle visits node , the value is 1; otherwise, the value is 0 is a 0-1 variable. If the alternative distribution center of the driverless vehicle is enabled, the value is 1; otherwise, the value is 0 is a 0-1 variable. If the vehicle drives from node to node , the value is 1; otherwise, the value is 0 It can be understood that in this embodiment, the distribution cost analysis includes Let be the cost per unit distance of the truck (unit: yuan / km). The driving distance cost of all trucks is (1) Let be the cost per unit time of the driverless vehicle (unit: yuan / minute). The usage time cost of all driverless vehicles is (2) Let be the cost per unit time of the driver (unit: yuan / minute). The labor cost includes the time cost of the truck driving, the time cost of the truck delivering customers and the distribution center. The labor cost of all drivers is (3) Let be the fixed usage cost when the candidate distribution center of the driverless vehicle is enabled for distribution (unit: yuan / unit). The usage cost of all enabled distribution centers is (4) 3 Mathematical Model With the goal of minimizing the total distribution cost, a location-routing model for the joint distribution of trucks and driverless vehicles is constructed as follows: (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) Equation (5) is the objective function, which minimizes the total distribution cost, that is, the sum of the driving distance costs of all trucks, the usage time costs of the driverless vehicles, the labor costs of the drivers, and the fixed usage costs of the selected collection and distribution points. Equation (6) means that each customer point can only be served by one vehicle once. Equation (7) means that when the collection and distribution point for the driverless vehicle is selected, the driverless vehicle can serve the customers within the collection and distribution area. Equation (8) means that when the collection and distribution point for the driverless vehicle is selected, the truck can make deliveries for this collection and distribution point. Equation (9) means that the driverless vehicle cannot cross multiple collection and distribution point areas for deliveries. Equation (10) means that the driverless vehicle cannot serve customers outside the collection and distribution point area. Equation (11) means that the flow of the driverless vehicle and the truck in and out of the node is conserved. Equation (12) means that the truck departs from the distribution center, completes the delivery, and then returns to the distribution center. Equation (13) means that the driverless vehicle departs from the collection and distribution point, completes the delivery, and then returns to the original collection and distribution point. Equation (14) means that each truck is enabled at most once. Equations (15) and (16) respectively represent the time calculation methods when the truck and the driverless vehicle arrive at the next node. Equation (17) means that the departure time of the vehicle at a node is equal to the sum of the arrival time of the node and the service time of the node. Equation (18) represents the change in the load capacity of the vehicle before and after serving a node. Equation (19) represents the calculation method for the remaining available driving distance when the driverless vehicle arrives at the customer point. Equation (20) means that the remaining driving distance after the driverless vehicle serves a customer must be able to return to the original collection and distribution point. Equation (21) represents the truck capacity constraint. Equation (22) means that the unloading volume of the truck at a collection and distribution point is equal to the sum of the demand volumes of all customers within the service area of this collection and distribution point. Equation (23) represents the driverless vehicle capacity constraint. Equation (24) is the value constraint of the decision variable.

[0064] It should be noted that the Vehicle Routing Problem (VRP) is a typical NP-Hard problem, and it is difficult to find the global optimal solution using traditional exact algorithms. The location-routing problem of combined distribution of trucks and autonomous vehicles is more complex and difficult to solve than the traditional VRP. The Adaptive Large Neighborhood Search algorithm (ALNS) is a heuristic algorithm with a flexible operation mechanism, self-adaptability, and efficient search ability, and it is widely used to solve various combinatorial optimization problems. Therefore, this paper designs an Improved Adaptive Large Neighborhood Search algorithm (IALNS) to solve the problem according to the characteristics of the model. The specific idea is as follows: (1) Considering the different customer demands faced by logistics enterprises in each distribution, a dynamic location strategy for the distribution collection and distribution points of autonomous vehicles is designed. (2) To expand the path search range and improve the quality of the solution, various destruction operators and repair operators are designed for the truck path. The destruction operators destroy the local structure of the current solution and break the original path restrictions, while the repair operators reconstruct the path after destruction; for the autonomous vehicle path, various neighborhood search operators are designed to enhance the optimization search ability of the algorithm. (3) To avoid the algorithm falling into local optimum, the simulated annealing acceptance criterion is introduced to accept worse solutions with a certain probability. By controlling the temperature parameter, the global search ability of the algorithm and the diversity of the solution are improved. The flowchart of IALNS is as Figure 5 shown.

[0065] In the specific implementation, the steps of the IALNS algorithm are as follows: Step1: Initialization.

[0066] Step1.1: Input data such as the coordinates of the distribution center, the coordinates of the candidate distribution collection and distribution points of autonomous vehicles, the service range of autonomous vehicles, the coordinates of customers, the capacity of trucks, the capacity of autonomous vehicles, the customer demand volume, and the initial selection threshold of the distribution collection and distribution points of autonomous vehicles.

[0067] Step1.2: Let be the maximum number of iterations, be the number of outer iterations, be the number of inner iterations, be the optimal total distribution cost, be the optimal path, be the initial temperature, be the temperature cooling coefficient, be the length of the Markov chain. Set , , , .

[0068] Step 2: Select the location of the distribution centers for driverless vehicles.

[0069] Step 2.1: According to the selection threshold select each distribution center for driverless vehicles to determine whether the threshold condition is met.

[0070] Step 2.2: Encode all the distribution centers for driverless vehicles to form a selection status coding sequence. Denote that the distribution center is selected as 1 and not selected as 0.

[0071] Step 2.3: Traverse the set of distribution center selection statuses, and determine whether the selection status coding sequence already exists. If it already exists, find the corresponding set with the same distribution center selection status, and use the optimized path result, operator weight, and annealing temperature variable in the inner layer as the initial variables, then go to Step 4.2. Otherwise, go to Step 3.

[0072] Step 3: Construct the initial path using the greedy strategy.

[0073] Step 3.1: For all the selected distribution centers , screen the customer points that meet the conditions . If , insert the customer point into the set of driverless vehicle deliveries , where is the distribution center for driverless vehicle deliveries and the customers are within the service range.

[0074] Step 3.2: Remove the customer points for driverless vehicle deliveries from all the sets of delivery customer points, and at the same time insert the selected distribution centers to construct the set of truck deliveries .

[0075] Step 3.3: Calculate the objective function value (the distance from the customer to the distribution center) for each customer in the set of driverless vehicle deliveries , and sort the customers in the set in ascending order of the objective function value.

[0076] Step 3.4: Select the customers from the sorted set of driverless vehicle deliveries in ascending order, and insert the customers into the path to be delivered. If the current delivery set is empty, go to Step 3.7. Otherwise, go to Step 3.5.

[0077] Step 3.5: When the customer point is inserted into the current path to be delivered, determine whether the capacity constraint is met. If it is met, go to Step 3.6. Otherwise, save the original path and add a new path for driverless vehicle deliveries.

[0078] Step3.6: Determine whether the remaining driving distance of the unmanned vehicle meets the power constraint when delivering to this customer. If it meets the constraint, save this customer point as the current delivery route of the unmanned vehicle and go to Step3.4. Otherwise, save the original route and add a new delivery route for the unmanned vehicle.

[0079] Step3.7 Calculate the distances between each customer in the truck delivery set and the distribution center, and sort the calculation results in ascending order.

[0080] Step3.8 Select customers from the set one by one in ascending order, and determine whether the capacity constraint is met. If it is met, go to Step3.9. Otherwise, save the original route and add a new delivery route for the truck.

[0081] Step3.9 Determine whether the current set is an empty set. If it is met, go to Step4. Otherwise, go to Step3.8.

[0082] Step4: Adaptive large neighborhood search.

[0083] Step4.1: Initialize the route, let the initial solution be , the current solution be , , initialize the weight coefficients of each operator to 1, and the scores and usage times of each operator to 0.

[0084] Step4.2: According to the operator weights, select one of the three destruction operators (random destruction, worst-cost destruction, related destruction) to destroy the truck route according to the roulette wheel method to obtain the solution .

[0085] (1) Random destruction operator: Determine the number of customers to be destroyed according to the total number of customers to be delivered , randomly destroy from the delivery tasks until customers are destroyed.

[0086] (2) Worst-cost destruction: Calculate the difference between the cost after each customer point in the route is destroyed and the cost of the original route that has not been destroyed, and destroy the results in descending order until the specified destruction quantity is reached.

[0087] (3) Related destruction: Calculate the similarity degree between two customer points according to formula (25), and remove them in descending order of similarity until the specified destruction quantity is reached.

[0088] (25) Among them, and are weight coefficients, is the distance from customer point to customer point , and are the demand volumes of customer points and customer point respectively. If customer point and customer point are on the same path is 0, otherwise it is 1.

[0089] Step4.3: According to the operator weights, select one of the three repair operators (random repair, greedy repair, regret repair) by roulette method to repair the truck path to obtain a solution .

[0090] (1) Random repair: Randomly select a customer point from the set of damaged customer points and insert it into a random position in the truck delivery path until there are no remaining customer points to insert.

[0091] (2) Greedy repair: Select a customer point from the set of damaged customer points, calculate the increase in delivery cost when the customer point is inserted into different positions of different truck delivery paths, and select the insertion position with the smallest increase in delivery cost and insert it into that position until there are no remaining customer points to insert.

[0092] (3) Regret repair: Calculate the delivery cost at the optimal insertion position and the delivery cost at the sub-optimal insertion position for each customer in the set of damaged customer points. The difference between the costs of the two positions is used as the regret value, and the customer point with the largest regret value is selected and inserted into the optimal position.

[0093] Step4.4: Select one of the four neighborhood search operators (Swap operator, Insert operator, 2-opt operator, or-opt operator) by roulette method to obtain the delivery path of the unmanned vehicle . Among them, the schematic diagram of the neighborhood operator is as Figure 6 shown.

[0094] (1) Swap operator: Swap the positions of two randomly selected customer points in the unmanned vehicle delivery path (as Figure 6 (a)); (2) Insert operator: Randomly select two customer points in the unmanned vehicle delivery path , and insert customer point after customer point (as Figure 6 (b)); (3) The 2-opt operator means randomly select 2 customer nodes , and reverse the customer positions between customer node and customer node (as Figure 6 (c)); (4) or-opt operator: Randomly select two customer points on the delivery route of the driverless vehicle , and insert the customer point and the customer point after flipping them together after the customer point (such as Figure 6 (d)); Step4.5: Calculate the sum of the delivery costs of the driverless vehicle and the truck, and update the routes of the driverless vehicle and the truck according to the calculation results. If , then let . On this basis, if , then let , . Otherwise, update according to the acceptance criterion. The acceptance criterion is as follows: When the new solution is worse than the current solution, accept it with the probability in Equation (26): (26) where is the cost of the new solution, is the cost of the current solution, is the annealing temperature of the current iteration.

[0095] Step4.6 Update the weights of each operator. The scores of the destruction operator and the repair operator depend on the solution. Specifically: (1) When , the operator score is increased by ; (2) When the condition in (1) is satisfied and , the operator score is increased by ; (3) When and the acceptance criterion is satisfied, the operator score is increased by ; (4) When and the acceptance criterion is not satisfied, the operator score is increased by ; After each iteration, the operator weights are updated according to Equation (27): (27) where and are the weights of the destruction and repair operators in the current iteration and the previous iteration respectively, is the operator decay coefficient, and are the operator score and the number of uses respectively.

[0096] Step4.7: Update the annealing temperature , update the iteration count , if , then go to Step 4.2; otherwise, output the optimal distribution cost , the optimal path , the current distribution cost , the current distribution path , the annealing temperature and the operator weights .

[0097] Step 5: Calculate the distribution cost of the unmanned vehicle distribution collection and distribution point and update the selection threshold.

[0098] Step 5.1: Determine whether the current collection and distribution point selection coding sequence already exists in the collection and distribution point selection status set. If it does not exist, create a new collection and distribution point selection status variable, and save the optimized distribution path , the annealing temperature and the operator weights . Otherwise, update the original distribution path, annealing temperature and operator weights.

[0099] Step 5.2: Calculate the optimized distribution path in the manner of Step 1 in the unmanned vehicle collection and distribution point location selection strategy.

[0100] Step 5.3: Traverse each selected unmanned vehicle distribution collection and distribution point and update its selection threshold according to Equation .

[0101] Step 6: End condition of IALNS.

[0102] , if , then go to Step 2; otherwise, the algorithm ends and outputs the result and .

[0103] In specific implementation, this embodiment proposes a distribution mode of combined distribution of trucks and unmanned vehicles, and realizes the optimization of the distribution network through the functional complementarity of the two transportation carriers. For high-density distribution scenarios such as urban communities, giving full play to the load capacity, endurance time and continuous operation of unmanned vehicles in terminal distribution helps to reduce the manual work intensity during distribution and improve the distribution efficiency.

[0104] Compared with the fixed hierarchical method of traditional hierarchical distribution, this embodiment proposes that the truck can distribute to the scattered customer points along the way when going to the collection and distribution point. This method is more flexible in the selection of distribution points. By slightly increasing the truck path, it replaces the longer path of unmanned vehicle distribution, effectively reducing the driving distance and expanding the path optimization space.

[0105] This embodiment proposes a dynamic location selection strategy for the distribution points of driverless vehicles based on distribution requirements. According to different customer distribution characteristics, it analyzes the structural changes in distribution requirements in the terminal area and adaptively selects the enabled status of distribution points. It effectively solves the problem of resource misallocation existing in the traditional fixed distribution point layout and realizes the best matching of facility resources and distribution requirements.

[0106] This embodiment designs an improved adaptive large neighborhood search algorithm, designs a variety of neighborhood operators to improve the search efficiency, and introduces the simulated annealing acceptance criterion to enhance the global search ability. It effectively optimizes the distribution routes of trucks and driverless vehicles and realizes the minimization of distribution costs.

[0107] In the specific implementation, a numerical example experiment was also carried out on the corresponding algorithm in the embodiment. 1. Experimental settings Since there is currently no standard test case library for the location-routing problem of combined truck and driverless vehicle distribution, the experiments in this paper generate test cases based on Solomon test cases. The generated test cases include three distribution types: clustered distribution (C1, C2), random distribution (R1, R2), and mixed distribution (RC1, RC2). Among them, the service time of customers and distribution points in each test case is set to 2 minutes. In addition, to meet the test requirements of this paper, the following data is supplemented: each test case uses coordinates , , and as candidate distribution points for driverless vehicle distribution, and the service area radius of each candidate distribution point is 12 km.

[0108] The model parameter settings are as follows: , , , , yuan / km, yuan / min, yuan / min, yuan. The IALNS parameter settings are as follows: , , , , , , , , , .

[0109] IALNS is programmed using Matlab 2023a and runs on a microcomputer with a CPU of 12th Gen Intel(R) Core(TM) i7-12700H@2.30 GHz and a RAM of 16GB.

[0110] 2 Experimental Analysis 2.1 Experiments with Multiple Types of Examples Experiments were conducted with different types of examples. The calculation results are shown in Table 1, where: IN represents the example name, TC represents the total distribution cost (unit: yuan), TT represents the total driving time of trucks and autonomous vehicles (unit: min), TRC represents the driving cost of trucks (unit: yuan), UC represents the usage cost of autonomous vehicles (unit: yuan), LC represents the labor cost (unit: yuan), DPC represents the fixed usage cost of the distribution collection points for autonomous vehicles (unit: yuan), UD represents the driving distance of autonomous vehicles (unit: km), DPN represents the number of selected collection points (unit: pieces), and RT represents the running time of the algorithm (unit: seconds).

[0111] According to the results in Table 1: (1) From the values of TC, TRC, UC, LC, and DPC, it can be obtained that the proportions of the truck driving cost, the cost of using the driverless vehicle, the labor cost, and the fixed cost of using the driverless vehicle distribution collection point in the total distribution cost are 68.28%, 4.93%, 21.41%, and 5.38% respectively. This shows that the truck driving cost and the labor cost are the main influencing factors of the logistics cost. Therefore, logistics enterprises should reasonably plan the driving routes of trucks, optimize the distribution mode, minimize the truck driving cost, reduce the number of customer points directly delivered by trucks, and control the labor cost of truck distribution. (2) From the values of TC, UC, and UD, it can be seen that the average driving mileage of the driverless vehicle reaches 245.33 km. Compared with the relatively high driving cost and labor cost of trucks, the driverless vehicle only consumes electric energy during the distribution process and does not require labor, and the distribution cost is significantly reduced, only accounting for 4.93% of the total distribution cost. This indicates that the use of driverless vehicle distribution is both economical and environmentally friendly. Therefore, logistics enterprises should give priority to driverless vehicle distribution, which not only helps to reduce the logistics cost, but also can improve the logistics distribution performance and promote the sustainable development of the logistics industry. (3) From the values of UC, DPC, and DPN, it can be known that the cost of using the driverless vehicle and the fixed cost of using the collection point increase with the increase in the number of selected collection points, indicating that the number of selected collection points will affect the total distribution cost within the collection point area when using driverless vehicle distribution. Therefore, logistics enterprises should reasonably select the driverless vehicle distribution collection points according to the actual distribution of customers to be served. (4) From the value of DPN, it can be seen that under different customer distribution types, the number of selected driverless vehicle distribution collection points is different. In the C-class example with clustered distribution, the number of selected driverless vehicle distribution collection points is the largest; in the R-class example with random distribution, the number of selected driverless vehicle distribution collection points is the smallest; in the RC-class example with mixed distribution, the number of selected driverless vehicle distribution collection points is between the R-class and C-class examples. This shows that the dynamic location selection strategy of the driverless vehicle distribution collection point designed in this paper can flexibly select locations according to different types of customer distribution, and has rationality and feasibility. (5) From the value of RT, it can be seen that the average running time of the algorithm is 370.63, which indicates that IALNS can solve large-scale examples of various types within a reasonable time and has high practicality and running efficiency.

[0112] Table 1 Experimental Results of Different Examples The location-routing diagrams of the driverless vehicle for Examples C101, C201, R101, and RC202 are as Figure 7 shown.

[0113] From Figure 7 it can be seen that: The driverless vehicle is responsible for delivering customers within the selected collection point area, and the truck is responsible for delivering customers at the selected collection points and within the unselected collection point areas.

[0114] Figure 7 (a) and Figure 7 In (a) and (b), the customers are in a clustered distribution, and the number of selected collection and distribution points is four in each case. Three trucks are used at the distribution centers respectively. In the C1 type of example, one driverless vehicle is used at each collection and distribution point. In the C2 type of example, only one collection and distribution point uses one driverless vehicle, and two driverless vehicles are used at the other three collection and distribution points. Figure 7 In (c), the customers are in a random distribution, the number of selected collection and distribution points is two, two trucks are used at the distribution center, and one driverless vehicle is used at each selected collection and distribution point. Figure 7 In (d), the customers are in a mixed distribution, and the number of selected collection and distribution points is three. Three trucks are used at the distribution center, and one driverless vehicle is used at each selected collection and distribution point. It shows that the dynamic site selection strategy for driverless vehicle distribution can, according to the changes in customer demand, compare the costs of truck distribution and driverless vehicle distribution in the collection and distribution point area, and carry out the dynamic site selection of the collection and distribution points for driverless vehicle distribution. When the customer demand in a certain collection and distribution point area is large and the cost of using a driverless vehicle is low, this collection and distribution point is preferentially selected for use; when the customer demand in a certain collection and distribution point area is small and the cost of using a truck is low, then this collection and distribution point is not selected. Figure 7 The site selection - path planning results of different types of examples show that the algorithm in this paper can dynamically select the sites of the collection and distribution points according to the customer distribution, and reasonably plan the paths of trucks and driverless vehicles, which is reasonable and feasible.

[0115] 2.2 Comparative experiments between the dynamic site selection strategy and the fixed site selection strategy Comparative experiments are carried out on the dynamic site selection strategy and the fixed site selection strategy using multiple types of examples (the fixed site selection strategy means that all the preset collection and distribution points for driverless vehicle distribution are enabled without considering customer demand). The experimental results are shown in Table 2. Among them, UCT represents the sum of the usage cost of the driverless vehicle and the fixed usage cost of the collection and distribution point (collectively referred to as the driverless vehicle distribution cost), UVN represents the number of driverless vehicles used, and NU represents the total number of customers in the area of the selected collection and distribution points for driverless vehicle distribution. The meanings of the other symbols are the same as those in Table 1.

[0116] It can be seen from the results in Table 2 that: (1) According to the results of TC, UCT, and DPN, the total distribution cost of the dynamic location strategy in the R type and RC type cases is significantly lower than that of the fixed location strategy, with an average savings of 3.76%, and the distribution cost of the driverless vehicle is saved by an average of 21.84%. This shows that the dynamic location strategy can scientifically enable the collection and distribution points to carry out distribution according to customer needs, avoid selecting unnecessary collection and distribution points, and effectively reduce the distribution cost. In contrast, the fixed location strategy fixedly selects all collection and distribution points. Even if the demand in some areas is small, driverless vehicles are still used for distribution, resulting in an increase in the number of distribution areas and high operating costs. (2) According to DPN and NU, it can be seen that in the C type cases, the customers are concentrated within each collection and distribution point area, and the number of selected collection and distribution points by the dynamic location strategy is the same as that of the fixed location strategy. This shows that in the case of concentrated customer distribution within the collection and distribution point area, the distribution cost of using driverless vehicles is lower. In the R type and RC type cases, when the dynamic location strategy is adopted, the number of customers served by driverless vehicles only slightly decreases compared with the fixed location strategy, and the number of selected collection and distribution points also decreases accordingly. This indicates that the number of customers in the unselected collection and distribution point areas is small, and the operating cost generated by continuing to use driverless vehicles is higher than that of truck delivery. Therefore, only some collection and distribution points with high customer demand are selected. (3) Generally speaking, the dynamic location strategy can respond more precisely to changes in customer needs, reasonably enable the collection and distribution points for driverless vehicle distribution, effectively reduce the operating costs of unnecessary collection and distribution points, and improve the resource utilization efficiency. Therefore, when logistics enterprises conduct location selection for collection and distribution points, they should adopt the dynamic location strategy to reasonably enable the collection and distribution points to carry out distribution according to the actual customer needs within the collection and distribution point area, ensure that the resource allocation matches the demand, effectively control the logistics cost, and improve the logistics distribution performance.

[0117] Table 2 Experimental Results of Dynamic Location Strategy and Fixed Location 2.3 Comparative Experiment between Driverless Vehicle Distribution and Traditional Vehicle Distribution Multi-type cases are adopted to conduct a comparative experiment on the combined distribution mode of trucks and driverless vehicles and the traditional vehicle distribution mode (traditional vehicle distribution means that all customers are delivered by trucks). The experimental results are shown in Table 3. Among them, TR represents the driving time of the truck, TS represents the manual service time, VN represents the number of trucks used, CSR represents the cost saving ratio of the total distribution cost under different distribution modes, and TSR represents the time saving ratio of the sum of the driving time of the truck and the manual service time under different distribution modes. The meanings of the remaining symbols are the same as those in Table 1.

[0118] Table 3 Experimental Results of Truck and Driverless Vehicle Distribution Modes and Traditional Vehicle Distribution Mode It can be seen from the experimental results in Table 3 that: (1) According to the results of TC and CSP, the total distribution cost under the combined distribution mode of trucks and driverless vehicles is significantly lower than that of the traditional vehicle distribution strategy. Among them, the total distribution cost is saved by up to 8.38% at most, 4.51% at least, and 7.49% on average. Since driverless vehicle distribution does not require labor costs and has a lower distribution cost, in the scenarios of short-distance distribution and end distribution where customer demands are concentrated, driverless vehicles can effectively replace traditional vehicles to complete distribution tasks and reduce the overall distribution cost. In contrast, due to the need to be equipped with manual driving and manual services, the operating cost of traditional vehicle distribution not only includes vehicle driving costs but also labor costs, showing obvious disadvantages in end distribution where customers need to be served frequently. Therefore, logistics enterprises should reasonably adopt driverless vehicle distribution on the basis of truck distribution in combination with the specific characteristics of the distribution scenario, make full use of the cost advantages of driverless distribution, and reduce the total distribution cost. (2) From the results of TR, TS, and TSR, when the truck and driverless vehicle distribution mode is adopted, the driving time of the truck and the manual service time are both significantly shorter than those of the traditional vehicle distribution mode. Specifically, the sum of these two times is saved by up to 24.69% in the C1 case and 15.53% in the R2 case at least, with an average saving of 19.03%. This indicates that the combined distribution mode of trucks and driverless vehicles can effectively reduce the redundant links in the logistics distribution process and reduce the use time of trucks by flexibly assigning tasks to driverless vehicles. In the case of high human resource costs, logistics enterprises adopting the combined distribution strategy of trucks and driverless vehicles can not only significantly reduce the dependence on manual services, reduce the operating pressure, but also further improve the logistics distribution performance.

[0119] 2.4 Algorithm Comparison Experiment To verify the effectiveness of the IALNS algorithm, a simulated annealing algorithm (SAA) for solving the location-routing problem of combined distribution of trucks and driverless vehicles was designed. The initial temperature of the SAA was set to 100, the cooling coefficient was 0.92, the length of the Markov chain was 300, and the number of iterations was 800. Comparative experiments were carried out using multiple types of examples, and the experimental results are shown in Table 4. Among them, TD represents the total driving distance of the vehicle, DSR represents the saving ratio of the total driving distance of the vehicle, and the meanings of the other symbols are the same as those in Table 1.

[0120] Table 4 Comparison Results of Different Optimization Algorithms From the experimental results in Table 4, it can be seen that: (1) According to the values of TC and TD, the total distribution cost and total driving distance obtained by IALNS in different types of examples are better than those of SAA. Among them, the total distribution cost is saved by up to 9.63%, at least 4.99%, and on average 6.84%. The total driving distance is saved by up to 5.28%, at least 1.88%, and on average 3.05%. This shows that IALNS can effectively reduce the distribution cost, shorten the distribution distance, and reduce the distribution time. (2) According to the value of RT, when IALNS solves the examples of type C and type RC, the solving time is significantly shorter than that of SAA. Although the solving time of IALNS is longer than that of SAA when solving the examples of type R, it is still within a reasonable time range, and generally, the time can be saved by an average of 7.08%. Therefore, IALNS is reasonable, reliable, and effective in solving the location-routing problem of combined truck and unmanned vehicle distribution.

[0121] 2.5 Sensitivity analysis (1) Sensitivity analysis of the service range of collection and distribution points To verify the influence of the service range of collection and distribution points on the path planning of trucks and unmanned vehicles, with other parameters unchanged, the service radius of collection and distribution points is changed in the range of [10, 18] with a step of 2, and the examples C206, R106, and RC106 are used for experiments. The experimental results are shown in Table 5, where: R represents the service radius of unmanned vehicles at collection and distribution points, UT is the driving time of unmanned vehicles, and the meanings of the remaining symbols are the same as those in Table 1 and Table 2.

[0122] As can be seen from the results in Table 5: (1) According to the values of TC, UT, and UCT, under different types of examples, as the service range of the collection and distribution points gradually expands, the total distribution cost shows an obvious downward trend, while the distribution time and cost of the unmanned vehicle show an upward trend. This is because the expansion of the service range of the collection and distribution points increases the number of customers served by the unmanned vehicle, and more distribution tasks are undertaken by the unmanned vehicle, while the distribution tasks of the truck are correspondingly reduced. Since the unmanned vehicle has lower driving costs and service costs, the reduction in the distribution cost of the truck is significantly greater than the increase in the distribution cost of the unmanned vehicle. Therefore, overall, the total distribution cost gradually decreases. In addition, as the service range of the unmanned vehicle expands, the unmanned vehicle needs to serve more customers, which inevitably leads to an increase in the driving time and distribution cost of the unmanned vehicle. (2) Different types of examples have different sensitivities to changes in the service range of the collection and distribution points. For the C2 and R1 examples: when R (the service radius of the unmanned vehicle at the collection and distribution point) increases from 10 to 16, the decline in the total distribution cost is significant, while when R increases from 16 to 18, the decline in the total distribution cost slows down significantly. For the RC1 example, when R increases from 14 to 16, the decline in the total distribution cost is relatively small, while in other intervals, the decline in the total distribution cost is more obvious. Therefore, when logistics enterprises expand the service range of the collection and distribution points, they should comprehensively consider customer needs and actual operating conditions, evaluate the necessity of expansion, and avoid additional distribution costs and time caused by excessive expansion of the service range.

[0123] Table 5 Experimental results of sensitivity analysis of the size of the collection and distribution area (2) Sensitivity analysis of the time cost of using unmanned vehicles To study the impact of the time cost of using unmanned vehicles on the location planning of unmanned vehicle collection and distribution points, with other parameters unchanged, the time cost of using unmanned vehicles varies in the range of [0.1, 0.5] with a step size of 0.1, and the example RC107 is used for the experiment. The experimental results are shown in Table 6, where: P represents the time cost of using unmanned vehicles, and the meanings of the other symbols are the same as those in Table 1. The experimental results are shown in Table 6, where: P represents the time cost of using unmanned vehicles, and the meanings of the other symbols are the same as those in Table 1.

[0124] Table 6 Experimental results of sensitivity analysis of the usage cost of unmanned vehicles As can be seen from the results in Table 6: (1) According to the values of P, UC, DPC, and DPN, as the time cost of using the unmanned vehicle increases, the number of selected distribution collection points for unmanned vehicle distribution and the fixed usage cost show a gradually decreasing trend. Because when the time cost of using the unmanned vehicle increases, the distribution collection point areas that originally had a distribution cost advantage gradually lose this advantage, resulting in the cost of using the unmanned vehicle for some distribution collection points being higher than that of truck distribution, so they are no longer selected. This indicates that under the same customer demand distribution, the selection of unmanned vehicle distribution collection points will be affected by the time cost of using the unmanned vehicle. (2) According to the values of P and TC, when P increases from 0.1 to 0.4, the total distribution cost shows an obvious upward trend; when P increases from 0.4 to 0.5, the total distribution cost increases relatively slowly. Therefore, in actual distribution planning, the time cost of using the unmanned vehicle should be fully considered, the number of enabled unmanned vehicle distribution collection points and the deployment ratio of unmanned vehicles should be reasonably selected, the distribution plan should be optimized, the total distribution cost should be reduced, and the distribution performance should be improved.

[0125] (3) Sensitivity analysis of the fixed usage cost of the distribution collection point To study the impact of the fixed usage cost of the distribution collection point on the location planning of unmanned vehicle distribution collection points, keeping other parameters unchanged, the fixed usage cost of a single distribution collection point is changed in the range of [30, 90] with a step of 15, and the example RC108 is used for experiments. The experimental results are shown in Table 7. Among them: Z represents the fixed usage cost of a single distribution collection point, and the meanings of the other symbols are the same as those in Table 1.

[0126] As can be seen from the results in Table 7: (1) According to the values of Z and DPN, as the fixed usage cost of a single distribution collection point increases, the number of selected distribution collection points decreases. Because the increase in the fixed usage cost of a single distribution collection point leads to an increase in the total distribution cost of the distribution collection point area. If the number of customers in the distribution collection point area is small, it is difficult to expand the cost advantage through unmanned vehicle distribution, resulting in a relatively high unmanned vehicle distribution cost for the entire distribution collection point area, and finally the originally selected distribution collection points are no longer selected. (2) According to the values of Z, TC, UC, and DPN, when Z increases from 45 to 60, although the unmanned vehicle usage cost decreases due to the decrease in the number of selected distribution collection points, the total distribution cost increases. Because the distribution collection point areas originally served by unmanned vehicles are switched to truck distribution, and although the truck distribution cost is lower than the unmanned vehicle distribution cost in this area, it is still higher than the unmanned vehicle usage cost, so the total cost increases. In summary, the fixed usage cost of the distribution collection point will affect the selection of the distribution collection point. Therefore, logistics enterprises should pay attention to the fixed usage cost of the distribution collection point when selecting a location. If the fixed usage cost of the distribution collection point is too high and the number of customers in the distribution collection point area is small, then this distribution collection point should not be selected.

[0127] Table 7 Experimental results of sensitivity analysis of the fixed usage cost of the distribution collection point In this embodiment, when a delivery task is received, the candidate distribution collection points for the unmanned vehicle and their service scopes are obtained according to the delivery task; the customer requirements are obtained according to the delivery task, and the threshold for selecting collection points is dynamically adjusted to determine whether each candidate collection point is enabled; for the enabled collection points, the combined delivery route of the truck and the unmanned vehicle is optimized by an improved adaptive large neighborhood search algorithm to minimize the total delivery cost; the total delivery cost, the optimal route, and the collection point selection status are calculated and output. For the delivery mode of the combined delivery of the truck and the unmanned vehicle, a dynamic location selection strategy for the unmanned vehicle collection points based on delivery requirements is proposed, which flexibly allocates delivery tasks, reduces the manual work intensity, and improves the delivery efficiency.

[0128] In addition, an embodiment of the present application also proposes a computer-readable storage medium, on which a program for location-path decision-making of combined delivery of trucks and unmanned vehicles is stored. When the program for location-path decision-making of combined delivery of trucks and unmanned vehicles is executed by a processor, the steps of the method for location-path decision-making of combined delivery of trucks and unmanned vehicles as described above are implemented.

[0129] Refer to Figure 8 , Figure 8 which is a structural block diagram of the first embodiment of the location-path decision-making device for combined delivery of trucks and unmanned vehicles of the present application.

[0130] As Figure 8 shown, the location-path decision-making device for combined delivery of trucks and unmanned vehicles proposed in the embodiment of the present application includes: A data acquisition module 10, configured to obtain candidate distribution collection points for the unmanned vehicle and their service scopes according to the delivery task when the delivery task is received; A judgment module 20, configured to obtain customer requirements according to the delivery task, and dynamically adjust the threshold for selecting collection points to determine whether each candidate collection point is enabled; An algorithm module 30, for the enabled collection points, optimizing the combined delivery route of the truck and the unmanned vehicle by an improved adaptive large neighborhood search algorithm to minimize the total delivery cost; A calculation module 40, configured to calculate and output the total delivery cost, the optimal route, and the collection point selection status, where the total delivery cost includes the truck driving cost, the unmanned vehicle usage time cost, the driver labor cost, and the fixed usage cost of the collection point.

[0131] It should be understood that the above is only for illustration and does not constitute any limitation to the technical solution of the present application. In specific applications, those skilled in the art can set according to needs, and the present application does not make any restrictions on this.

[0132] In this embodiment, when receiving a delivery task, the candidate collection and distribution points for the unmanned vehicle delivery and their service scopes are obtained according to the delivery task; the customer requirements are obtained according to the delivery task, and the threshold for selecting collection and distribution points is dynamically adjusted to determine whether each candidate collection and distribution point is enabled; for the enabled collection and distribution points, the combined delivery route of the truck and the unmanned vehicle is optimized through an improved adaptive large neighborhood search algorithm to minimize the total delivery cost; the total delivery cost, the optimal route, and the status of collection and distribution point selection are calculated and output. A dynamic location strategy for unmanned vehicle collection and distribution points based on delivery requirements is proposed for the delivery mode of combined truck and unmanned vehicle delivery, which flexibly allocates delivery tasks, reduces the manual work intensity, and improves the delivery efficiency.

[0133] It should be noted that the above-described work process is only illustrative and does not limit the protection scope of this application. In actual applications, those skilled in the art can select some or all of them according to actual needs to achieve the purpose of the solution of this embodiment, and no limitation is imposed here.

[0134] In addition, for the technical details not described in detail in this embodiment, reference can be made to the location-routing decision method for combined truck and unmanned vehicle delivery provided in any embodiment of this application, which will not be elaborated here.

[0135] In addition, it should be noted that in this article, the terms "including", "comprising", or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or system including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article, or system. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article, or system including that element.

[0136] The serial numbers of the embodiments of the present application above are only for description and do not represent the superiority or inferiority of the embodiments.

[0137] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases, the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as Read-Only Memory (ROM) / RAM, magnetic disk, optical disk), and includes several instructions to enable a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of various embodiments of the present application. The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A location-routing decision-making method for combined distribution of trucks and autonomous vehicles, characterized in that Including: When receiving a delivery task, obtaining the candidate distribution collection points and their service scopes for the unmanned vehicle according to the delivery task; Obtaining customer requirements according to the delivery task, and dynamically adjusting the collection point selection threshold to determine whether each candidate collection point is enabled; For the enabled collection points, optimizing the combined delivery path of the truck and the unmanned vehicle through an improved adaptive large neighborhood search algorithm to minimize the total delivery cost; Calculating and outputting the total delivery cost, the optimal path, and the collection point selection status, where the total delivery cost includes the truck driving cost, the unmanned vehicle usage time cost, the driver labor cost, and the fixed usage cost of the collection point.

2. The method according to claim 1, wherein The specific process of dynamically adjusting the collection point selection threshold includes: Calculating the difference between the unmanned vehicle delivery cost when the collection point is enabled and the truck delivery cost when the collection point is not enabled; Dynamically updating the selection threshold according to the difference. If the unmanned vehicle delivery cost is lower, increasing the selection threshold to increase the probability of the collection point being enabled; otherwise, decreasing the selection threshold; Judging whether the collection point is enabled through the normal distribution function, and generating a collection point selection status coding sequence.

3. The method according to claim 1, characterized in that The improved adaptive large neighborhood search algorithm includes: Using a destruction operator to locally destroy the truck path, including random destruction, worst-cost destruction, and related destruction; Using a repair operator to reconstruct the destroyed path, including random repair, greedy repair, and regret repair; Designing a neighborhood search operator for the unmanned vehicle path, including Swap, Insert, 2-opt, and or-opt operators; Introducing a simulated annealing acceptance criterion to accept a worse solution in a probabilistic manner to avoid the algorithm falling into a local optimum.

4. The method according to claim 3, wherein The similarity calculation of the destruction operator is: Among them, and are weight coefficients, is the distance from customer point to customer point . and are respectively the demand quantities of customer point and customer point . If customer point and customer point are on the same path, is 0, otherwise it is 1. Remove them in descending order of similarity until the specified destruction quantity is reached.

5. The method according to claim 3, wherein The steps of using the repair operator to reconstruct the destroyed path include: Obtain the operator weights and, according to the operator weights, select one of the three repair operators by roulette method to repair the truck route to obtain a solution ; Among them, random repair includes: randomly selecting a customer point from the set of destroyed customer points, and inserting it into a random position of the truck delivery path until there are no remaining customer points to insert; Greedy repair includes: selecting a customer point from the set of destroyed customer points, calculating the delivery cost increment when the customer point is inserted into different positions of different truck delivery paths, and selecting the insertion position with the smallest delivery cost increment and inserting it into that position until there are no remaining customer points to insert; Regret repair includes: calculating the delivery cost at the optimal insertion position and the sub-optimal insertion position for each customer in the set of destroyed customer points, and taking the difference between the costs of the two positions as the regret value, and selecting the customer point with the largest regret value and inserting it into the optimal position.

6. The method according to claim 3, wherein The steps of designing a neighborhood search operator for the unmanned vehicle path include: Select one from the four neighborhood search operators according to the roulette method to obtain the path of the distribution unmanned vehicle ; Among them, the Swap operator exchanges the positions of two randomly selected customer points in the unmanned vehicle delivery path; The Insert operator randomly selects two customer points on the delivery route of the driverless vehicle , and inserts the customer point after the customer point ; The 2-opt operator randomly selects two customer nodes and reverses the customer positions between customer node and customer node The or-opt operator randomly selects two customer points on the delivery route of the driverless vehicle , and and customer point are flipped and inserted together after customer point .

7. The method according to claim 1, characterized in that, The mathematical model of the total delivery cost is: Among them, is the cost of the driving distance of the truck, is the cost of the usage time of the driverless vehicle, is the labor cost of the driver, is the fixed usage cost of the collection and distribution point, and satisfies the path flow conservation and capacity constraints.

8. A location-routing decision-making device for combined distribution of trucks and autonomous vehicles, characterized in that Executing the method according to claim 1, including: A data acquisition module, used for obtaining the candidate distribution collection points and their service scopes for the unmanned vehicle according to the delivery task when receiving the delivery task; A judgment module, used for obtaining customer requirements according to the delivery task, and dynamically adjusting the collection point selection threshold to judge whether each candidate collection point is enabled; The algorithm module optimizes the combined distribution path of trucks and driverless vehicles for the enabled collection and distribution points through an improved adaptive large neighborhood search algorithm to minimize the total distribution cost. The calculation module is used to calculate and output the total distribution cost, the optimal path, and the selection status of the collection and distribution points. Among them, the total distribution cost includes the truck driving cost, the driverless vehicle usage time cost, the driver labor cost, and the fixed usage cost of the collection and distribution points.

9. A computer device, characterized in that, The device includes: a memory and a processor. When the processor runs the computer instructions stored in the memory, it executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It includes instructions that, when run on a computer, cause the computer to execute the method according to any one of claims 1 to 7.

Images