Parcel delivery method, device and equipment based on unmanned aerial vehicle logistics system and storage medium

Abstract

The application discloses a package delivery method and device based on a UAV logistics system, equipment and a storage medium. The UAV logistics system comprises at least one set of heterogeneous UAVs and a plurality of delivery sites. The method comprises the following steps: constructing a task scheduling model; the task scheduling model takes minimizing the maximum completion time of all package delivery tasks as an objective function, and constructs a multi-dimensional constraint condition set based on the heterogeneity of the UAVs, the flight time limit and the load capacity constraint; a genetic algorithm is used to solve the task scheduling model to generate a high-quality task scheduling scheme that meets the multi-dimensional constraint condition set; and the at least one set of heterogeneous UAVs are controlled to perform package loading, path planning and site access according to the task scheduling scheme to complete the delivery task. Therefore, the task scheduling model and the genetic algorithm cooperate with each other to provide a targeted high-quality task scheduling scheme according to the actual situation, so that the resource utilization rate and the overall delivery efficiency are improved.

Description

Parcel delivery methods, devices, equipment, and storage media based on drone logistics systems Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) control technology, and more specifically, to a parcel delivery method, apparatus, equipment, and storage medium based on an UAV logistics system. Background Technology

[0002] With the rapid development of drone technology, drone logistics systems have demonstrated high efficiency and flexibility in scenarios such as express delivery and emergency supplies delivery. However, in practical applications, how to efficiently schedule multiple heterogeneous drones to achieve rapid completion of package delivery tasks remains a key issue that urgently needs to be addressed.

[0003] In existing technologies, task scheduling methods typically employ static allocation or simple heuristic rules (such as nearest neighbor method or greedy algorithm). These methods often lead to uneven task allocation, with some drones being overloaded or exceeding flight time limits, while other drones return to base early, resulting in low resource utilization and limited overall delivery efficiency. Summary of the Invention

[0004] In view of the above problems, this application proposes a parcel delivery method, apparatus, equipment and storage medium based on a drone logistics system, which can effectively solve the above problems.

[0005] In a first aspect, embodiments of this application provide a parcel delivery method based on an unmanned aerial vehicle (UAV) logistics system. The UAV logistics system includes at least one group of heterogeneous UAVs and multiple delivery stations. The method includes: constructing a task scheduling model; using the objective function of minimizing the maximum completion time of all parcel delivery tasks, and constructing a multi-dimensional constraint set based on the heterogeneity of the UAVs, flight time limitations, and payload capacity constraints; solving the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set; and controlling at least one group of heterogeneous UAVs to perform parcel loading, path planning, and station access according to the high-quality task scheduling scheme to complete the delivery task.

[0006] Secondly, embodiments of this application also provide a parcel delivery device based on an unmanned aerial vehicle (UAV) logistics system. This device is applied to an UAV logistics system comprising at least one group of heterogeneous UAVs and multiple delivery stations. The device includes: a construction module for constructing a task scheduling model; the task scheduling model uses minimizing the maximum completion time of all parcel delivery tasks as its objective function, and constructs a multi-dimensional constraint set based on the heterogeneity of the UAVs, flight time limitations, and payload capacity constraints; a generation module for solving the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set; and a delivery module for controlling at least one group of heterogeneous UAVs to perform parcel loading, path planning, and station access according to the high-quality task scheduling scheme to complete the delivery task.

[0007] Thirdly, embodiments of this application also provide a parcel delivery device, including a processor, a memory, and one or more applications; the one or more applications are stored in the memory and configured to be executed by the processor to implement the above-described parcel delivery method based on the drone logistics system.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing program code, wherein the above-described package delivery method based on the drone logistics system is executed when the program code is run by a processor.

[0009] The technical solution provided in this application includes a drone logistics system comprising at least one group of heterogeneous drones and multiple delivery stations. The package delivery method based on this drone logistics system includes: constructing a task scheduling model; the task scheduling model uses minimizing the maximum completion time of all package delivery tasks as the objective function, and constructs a multi-dimensional constraint set based on the heterogeneity of the drones, flight time limitations, and payload capacity constraints; using a genetic algorithm to solve the task scheduling model to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set; and controlling at least one group of heterogeneous drones to perform package loading, path planning, and station visits according to the task scheduling scheme to complete the delivery task. Thus, by combining the task scheduling model and the genetic algorithm, targeted high-quality task scheduling schemes can be provided based on actual conditions, thereby improving resource utilization and overall delivery efficiency. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments and drawings obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0011] Figure 1 shows a schematic diagram of the structure of an unmanned aerial vehicle (UAV) logistics system according to an embodiment of this application.

[0012] Figure 2 shows a flowchart of a parcel delivery method based on an unmanned aerial vehicle (UAV) logistics system provided in an embodiment of this application.

[0013] Figure 3 shows a schematic diagram of the structure of a parcel delivery device based on an unmanned aerial vehicle (UAV) logistics system provided in an embodiment of this application.

[0014] Figure 4 shows a schematic diagram of the structure of a parcel delivery device provided in an embodiment of this application.

[0015] Figure 5 shows a schematic diagram of the structure of a computer-readable storage medium provided in an embodiment of this application. Detailed Implementation

[0016] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0017] With the rapid development of drone technology, drone logistics systems have demonstrated high efficiency and flexibility in scenarios such as express delivery and emergency supplies delivery. However, in practical applications, how to efficiently schedule multiple heterogeneous drones to achieve rapid completion of package delivery tasks remains a key issue that urgently needs to be addressed.

[0018] In existing technologies, task scheduling methods typically employ static allocation or simple heuristic rules (such as nearest neighbor method or greedy algorithm). These methods often lead to uneven task allocation, with some drones being overloaded or exceeding flight time limits, while other drones return to base early, resulting in low resource utilization and limited overall delivery efficiency.

[0019] To address the aforementioned issues, this application provides a parcel delivery method, apparatus, device, and storage medium based on an unmanned aerial vehicle (UAV) logistics system. This method, applied to an UAV logistics system comprising at least one group of heterogeneous UAVs and multiple delivery stations, includes: constructing a task scheduling model; using the objective function of minimizing the maximum completion time of all parcel delivery tasks, and constructing a multi-dimensional constraint set based on the heterogeneity of UAVs, flight time limitations, and payload capacity constraints; solving the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set; and controlling at least one group of heterogeneous UAVs to perform parcel loading, path planning, and station access according to the task scheduling scheme to complete the delivery task.

[0020] Therefore, by combining the task scheduling model with the genetic algorithm, targeted and high-quality task scheduling solutions can be provided based on the actual situation, thereby improving resource utilization and overall delivery efficiency.

[0021] This application provides a drone logistics system, which includes at least one group of heterogeneous drones. and multiple delivery stations Heterogeneous drones are used to deliver different packages From the originating delivery station to the corresponding destination station, different heterogeneous drones in the drone logistics system have different flight speeds, payload capacities, and maximum flight times.

[0022] For example, please refer to Figure 1, which shows a structural schematic diagram of a drone logistics system according to an embodiment of this application. As shown in Figure 1, the drone logistics system includes at least one group of heterogeneous drones. and drones Multiple delivery stations serve as the originating delivery stations. Delivery stations Delivery stations Delivery stations and delivery stations The packages to be delivered are respectively packages ,pack ,pack ,pack and packages Drones Responsible for delivering the package From the originating delivery station Delivery to delivery station and the package From the originating delivery station Delivery to delivery station drones Responsible for delivering the package ,pack and packages From the originating delivery station Delivery and distribution stations Delivery stations Delivery stations and delivery stations In the corresponding site.

[0023] This application addresses the heterogeneous drone scheduling problem in drone logistics systems, aiming to minimize the maximum completion time of package delivery tasks. The heterogeneous drones will be located at the originating delivery station. A large number of parcels in the (i.e., transit station) are delivered to their destinations in the nearby area.

[0024] In the process of route planning for heterogeneous drones to deliver packages, drone logistics systems face significant challenges in drone scheduling due to the different flight speeds, payload capacities (which limit the number of packages a drone can deliver in a single flight), and maximum flight times (which limit the distance a drone can fly in a single flight). Specifically:

[0025] First, drones are heterogeneous; second, packages loaded on the same drone may have different destinations, so the docking sequence during flight must be determined; finally, due to the originating delivery station... The number of packages is enormous, and each drone may need to perform multiple flights, so the flight sequence for each drone must be determined.

[0026] In this application, a classical genetic algorithm is employed ( This technology aims to solve the aforementioned problems by effectively addressing the scheduling issues of drones in drone logistics systems, and it demonstrates significant advantages in global search capabilities. Specifically:

[0027] Please refer to Figure 2, which shows a flowchart of a package delivery method based on a drone logistics system according to an embodiment of this application, applicable to the aforementioned drone logistics system. As shown in Figure 2, the package delivery method based on the drone logistics system may include steps 110 to 130.

[0028] In step 110, a task scheduling model is constructed.

[0029] The task scheduling model takes minimizing the maximum completion time of all package delivery tasks as its objective function, and constructs a multi-dimensional set of constraints based on the heterogeneity of drones, flight time limits, and payload capacity constraints.

[0030] Flight time limits and payload capacity constraints can be determined by the status reporting data from the drone. In some implementations, this package delivery method based on a drone logistics system may further include the following steps:

[0031] (1) Receive periodic status reports from each UAV;

[0032] (2) Determine flight time limits and load capacity constraints based on status reporting data.

[0033] The status reporting data includes at least one of the following: current geographic location coordinates, flight altitude, flight speed, remaining battery power, current payload weight, onboard package identifier list, and progress information of executed tasks. In the embodiments of this application, the status reporting data includes the UAV's current location, current payload weight, and maximum permissible payload.

[0034] The flight time limit of the drone is determined by calculating the real-time distance between the drone's current location and the target delivery station; the payload capacity constraint of the drone is determined by comparing the current payload weight with the maximum allowable payload.

[0035] There is a station in the drone logistics system. As the starting point for delivery (i.e., the transfer station), and Each delivery station is distributed as a parcel locker. Any delivery station and delivery stations The distance between them is used express.

[0036] Assuming all packages requiring delivery are placed at the originating delivery station. The set of packages can be represented as: Any package in the package set The weight and destination are marked as follows: and ,and These packages will be delivered by a fleet of drones: Delivery to destination All drones were initially at the starting delivery station. Available. Drones. Its payload capacity, flight speed, and maximum flight time are respectively expressed as follows: , and express.

[0037] Package delivery is considered a delivery task. The drone logistics system is responsible for generating scheduling solutions for these tasks. The solution to complete all tasks can be represented as a set of solutions for each drone, i.e. . Indicates drone The solution to be implemented consists of flight missions, namely... .

[0038] During each flight, the drone It may stop at several delivery stations and unload some packages at each station. Therefore, a single flight can also be represented as a set of delivery stations, i.e. Drones The The first flight The station is represented as a triple. ,in, It is in time At the delivery station The collection of packages to be unloaded.

[0039] In one solution, each package should be assigned to a specific drone with a predetermined completion time, determined by the delivery station it is loaded at and the delivery station it is unloaded from. The drone logistics system aims to find a feasible solution that minimizes the completion time of all tasks.

[0040] For example, suppose there are four delivery stations and two drones, with the two drones having payloads of respectively... =2 kg =5 kg, average speeds are respectively =60 km / h =48 km / h, maximum flight time are respectively =0.5 hours =0.67 hours. Initially at the starting delivery station. Five packages were stored. , ..., The weights are respectively =1.5 kg =1.1 kg =0.7 kg =2 kg =2.5 kg. Their destinations are divided into = , = , = , = , = . Indicates delivery station and The distance between them.

[0041] Based on the above background, two feasible solutions were generated. and Assuming in In China, drones Packages were delivered via two flights. and packages drones Packages were delivered sequentially on a single flight. ,pack and packages .exist In China, drones Packages were delivered via two flights. ,pack and packages drones Packages delivered in a single flight and packages . It will take 6 minutes to complete all tasks. The total completion time is 3.5 minutes. Therefore, Superior ,Will As a high-quality task scheduling solution.

[0042] Based on the above description and assumptions, it is constructed as a task scheduling model. In a specific implementation, the objective function is:

[0043]

[0044] in," "The number of drones," "For drones" Number of flights, "For drones" In the The number of sites visited during the flight, "For drones" In the During the second flight, the visit to the The time at each delivery station, " is a binary variable, which is a wrapper Whether it was unloaded during that flight.

[0045] In some implementations, the multi-dimensional constraint set includes at least one or more of the following: flight time sequence constraints, return time constraints to the originating delivery station, initial time constraints, initial location constraints, package weight and drone payload capacity matching constraints, package destination and drone flight path matching constraints, single flight payload limit constraints, single flight time limit constraints, package allocation integrity constraints, and domain constraints of decision variables.

[0046] In one specific implementation, the multi-dimensional constraint set includes flight time sequence constraints, return time constraints to the originating delivery station constraints, initial time constraints, initial location constraints, package weight and drone payload capacity matching constraints, package destination and drone flight path matching constraints, single flight payload limit constraints, single flight time limit constraints, package allocation integrity constraints, and domain constraints for decision variables.

[0047] In one specific implementation, the flight time sequence constraint is as follows:

[0048] .

[0049] in," "For drones" In the During the second flight, the visit to the The time at each delivery station, "For delivery stations" to delivery station distance, "For drones" Flight speed;

[0050] In one specific implementation, the time constraint for returning to the originating delivery station is:

[0051] .

[0052] in," "For drones" In the The start time of the next flight (i.e., the time of departure from the originating delivery station), "For drones" In the The last stop on the flight (i.e., the...) Arrival time at each stop), "For delivery stations" To the originating delivery station The distance;

[0053] In one specific implementation, the initial time constraint is:

[0054] ;

[0055] in," "For drones" The start time of the first flight;

[0056] In one specific implementation, the initial position constraint is:

[0057] .

[0058] In one specific implementation, the constraint condition for matching the package weight with the drone's payload capacity is:

[0059] .

[0060] in," "for the first" The weight of the package, "For drones" Maximum load capacity, "For all packages," There are one or more drones Its maximum load capacity meets the conditions. "For the package" weight No more than drones Maximum load capacity ;

[0061] In one specific implementation, the constraint condition for matching the package destination with the drone flight path is:

[0062] .

[0063] in," "For the package" The final delivery location, "For drones" In the The first flight The location of each stop;

[0064] In one specific implementation, the single-flight payload limit constraint is as follows:

[0065] .

[0066] In one specific implementation, the single flight time limitation constraint is as follows:

[0067] .

[0068] in," "For drones" In the The start time of the next flight, "For drones" Maximum flight time.

[0069] In one specific implementation, the package allocation integrity constraint is as follows:

[0070] .

[0071] In one specific implementation, the domain constraint condition for the decision variable is:

[0072] .

[0073] In determining the drone flight number (i.e., the...) In the case of the second flight and the unloading dock for delivering packages, the first The number of drones can be calculated using the objective function. The flight time sequence constraint is a recursive formula for calculating the delivery time of the packages. The flight start time can be recursively calculated using the formula to return to the starting delivery station time constraint and the initial time constraint. The initial position constraint indicates that the last stop of the drone flight is the transfer station, meaning that each drone should return. The package weight matching constraint with the drone's payload capacity means that no package weight exceeds the payload capacity of any drone. The package destination matching constraint with the drone's flight path ensures that packages are delivered to their destinations. The single flight payload limit constraint restricts the total weight of packages loaded in a single flight from exceeding the drone's payload capacity. The single flight time limit constraint restricts the total duration of a single flight from exceeding the maximum flight time of the corresponding drone. The package allocation integrity constraint and the domain constraint of the decision variables ensure that each package can be delivered. Further:

[0074] In step 120, a genetic algorithm is used to solve the task scheduling model to generate a high-quality task scheduling scheme that satisfies a set of multi-dimensional constraints.

[0075] The genetic algorithm consists of initial population generation, fitness evaluation, and genetic operations. These genetic operations mainly include selection, crossover, and mutation.

[0076] A high-quality task scheduling scheme is obtained by solving the task scheduling model using a genetic algorithm to minimize the completion time of all delivery tasks. Specifically, in some implementations, the step "solving the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies a set of multi-dimensional constraints" may include the following steps:

[0077] (1) Encode the potential solution to the task scheduling problem as a chromosome; the chromosome is in one-dimensional arrangement form, representing the access priority sequence of all tasks to be delivered or destination sites;

[0078] (2) During the decoding process, based on the priority sequence in the chromosome, combined with the maximum payload capacity and maximum flight time constraints of each UAV, a feasible task allocation scheme that satisfies the payload and flight time constraints of the UAV is generated.

[0079] (3) Based on the feasible task allocation scheme, determine the complete flight path of the UAV and its corresponding task completion time;

[0080] (4) Based on the objective function, the maximum value in the task completion time is taken as the fitness value of the corresponding potential solution;

[0081] (5) During the population evolution process, selection operations are performed based on fitness values, combined with crossover and mutation operations, to generate a new generation of population;

[0082] (6) After multiple iterations, a high-quality task scheduling scheme is obtained.

[0083] In the embodiments of this application, chromosomes are represented in a two-dimensional matrix form, and the chromosomes are composed of... The sequence consists of several queues, where each queue is a sequence of packages with the same destination. The matrix contains... Line, corresponding There are several destination stations. The size of any row in the two-dimensional matrix is ​​equal to the number of packages at the corresponding destination. To further simplify the representation, a chromosome can be represented by a series of queues, i.e. , ,…, Each queue is assigned a destination and a series of packages to that destination.

[0084] Each chromosome is represented as a series of package allocation queues. In reality, the order of the packages in the allocation queues is determined by the decoding method applied when decoding the chromosome (i.e., the decoding method determines the specific order of the packages in the allocation queues, i.e., the order in which drones should deliver packages to the various delivery stations). Therefore, a chromosome can be simply represented as a series of destination stations, each delivery station corresponding to one or more packages, and the order in which these delivery stations are accessed is determined by the decoding method.

[0085] For example, two chromosomes and Two chromosomes and Packages on the same line have the same destination site.

[0086] ;

[0087] ;

[0088] in," "This is the first package," "This is the second package," "For the third package,..." "This is the 9th package."

[0089] To generate a feasible task allocation scheme, a decoding method is used to allocate packages to drones based on the priorities given by the chromosome. Specifically, in some implementations, the step "during the decoding process, allocating packages to drones according to the priority sequence represented by the chromosome to generate a feasible task allocation scheme that satisfies the drone's payload and flight time constraints" may include the following steps:

[0090] (1) Determine the priority of the UAV based on its return time and load capacity, and arrange the UAVs according to their priority to form a UAV priority queue;

[0091] (2) Convert the chromosome in the form of a two-dimensional matrix into multiple parcel allocation queues by row; each row in the parcel allocation queue corresponds to a destination station, and the order of the parcels in each row indicates their delivery priority at that station;

[0092] (3) In each roulette wheel selection strategy, the target package allocation queue is selected from multiple package allocation queues, and the available drone with the highest current priority is selected from the drone priority queue for task allocation;

[0093] (4) Repeatedly execute the roulette wheel selection strategy to dynamically allocate tasks in the target package allocation queue to the selected UAV and generate a feasible task allocation scheme.

[0094] The priority of a drone is a tuple consisting of two values: return time and payload capacity. Return time refers to the time it takes for the drone to complete its assigned task and return to the originating delivery station. The time (usually set to zero initially). If both drones have the same payload capacity, the drone that returns earlier has higher priority.

[0095] The roulette wheel selection strategy is executed repeatedly until all tasks in the package allocation queue have been assigned. For example, in the... In each iteration, if the package allocation queue is not empty, it is selected as the target package allocation queue. Then, some or all of the packages in the target queue are allocated to the highest-priority drone in the drone priority queue. It's important to note that if all packages in the target allocation queue have been allocated to drones, and the drone still has remaining loading space after completing its current package loading, it continues to select packages from other non-empty package allocation queues for loading until it can no longer accommodate any more packages. Once a drone completes a flight mission and is fully loaded, its priority is updated based on its estimated return time, and it is reinserted into the drone priority queue to participate in subsequent mission allocations.

[0096] Based on the start and end times and path information of each flight mission in the generated feasible task allocation scheme, the maximum completion time of all missions is calculated, and the maximum completion time is returned as the fitness value of the individual. This fitness value is used to measure the overall execution efficiency of the feasible task allocation scheme; the smaller the fitness value, the shorter the completion time of all missions. The genetic algorithm performs selection, crossover, and mutation operations based on the fitness values ​​of each individual in the population, prioritizing the retention of individuals with higher fitness, thereby guiding the population to evolve in the direction of reducing the maximum completion time and gradually approaching a high-quality task scheduling scheme.

[0097] After obtaining a feasible task allocation scheme, the task sequence for each UAV is determined according to the scheme, and the complete flight path and corresponding task completion time are calculated based on the inter-site distance matrix and flight speed parameters. Fitness values ​​are used to measure the quality of each chromosome. During population evolution, selection operations are performed based on fitness values, combined with crossover and mutation operations, to generate a new generation of the population. Specifically:

[0098] First, since the initial population is the starting point for iterative optimization by the genetic algorithm, its diversity and the quality of the solutions significantly affect the convergence speed of the genetic algorithm and the performance of high-quality task scheduling schemes. Therefore, this application provides a method for obtaining an initial population with high diversity and quality. Specifically, three generation methods are used collaboratively to generate the initial population. These three methods include: a Traveling Salesman Problem (TSP)-based method (for generating more reasonable path plans), a Local Search (LS) method (for quickly finding some good local optima), and a random method (for randomly generating chromosomes as part of the initial population to ensure population diversity).

[0099] In one specific implementation, the traveling salesman problem method, the local search method, and the random method contribute 20%, 40%, and 40% to the initial population, respectively. Specifically:

[0100] The distance matrix (calculated based on the geographical coordinates of all stations, representing the flight distance or time between any two stations) is used to model the Traveling Salesman Problem. Simulated Annealing (SA) optimizes the randomly generated initial station visit sequence to find a feasible route that visits each station exactly once and has the shortest total path. The station order corresponding to this route is encoded into a chromosome, called the TSP chromosome, which constitutes 20% of the initial population for the genetic algorithm.

[0101] In the local search method, the TSP chromosome is used as the initial sequence, and a weighted longest match (WLM) decoding method is performed on it to generate a specific task allocation scheme (i.e., the initial solution), which serves as the starting point for the local search. Subsequently, a greedy strategy is used to iteratively improve the initial solution: at each step, a neighborhood operation that can reduce the fitness value (such as the maximum completion time) is selected until no better solution can be found or the number of iterations exceeds a preset threshold.

[0102] The goal of local search is to improve the quality of the initial solution, making it as close as possible to the local optimum. This aligns with the overall goal of the genetic algorithm in optimizing task scheduling schemes—namely, minimizing the maximum completion time of all tasks. Each independent run of the local search yields a best solution corresponding to a sequence of site visits, called a local search chromosome. Through multiple runs, a set of high-quality local search chromosomes is generated, collectively forming 40% of the initial population of the genetic algorithm.

[0103] By combining the initial 40% individuals generated by random methods, a complete initial population with both diversity and initial quality is finally formed, which helps the genetic algorithm converge to a high-quality task scheduling scheme more quickly in the subsequent evolution process.

[0104] Thus, we can see that the initial population of the genetic algorithm is generated using a local search method. By selecting TSP chromosomes as the initial sequence, the WLM decoding method is used to generate initial solutions, and the algorithm is iteratively improved in a greedy manner. Finally, a series of high-quality local search chromosomes are obtained. These chromosomes will serve as the initial population members of the genetic algorithm, helping the algorithm converge to a high-quality task scheduling scheme more quickly.

[0105] In the process of population evolution, selection operations are performed based on individual fitness values. Combined with crossover and mutation, a new generation of the population is generated. Through multiple generations of iteration, the task scheduling scheme is gradually optimized, resulting in a high-quality task scheduling scheme. The selection operation is the first step in the genetic algorithm. It is used to select "excellent" individuals from the current population as parents to participate in subsequent crossover and mutation operations. Specifically, in the selection operation, two chromosomes are randomly selected from the current population. and The strategy employed is that the better the fitness, the higher the probability that the chromosome will be selected.

[0106] For example, suppose It is a chromosome. Adaptability, It is a chromosome. The worst fitness value. Then the chromosome... The probability of being selected can be expressed as:

[0107] ;

[0108] in," "For from chromosomes" Randomly select a subsequence from the given information.

[0109] The selection operation randomly selects two chromosomes from the current population. and Then, the crossover operation is based on chromosomes. and Generate offspring, specifically:

[0110] (1) From chromosomes Randomly select a subsequence From chromosomes Randomly select a subsequence . subsequence with sequence They are the same length.

[0111] (2) Use subsequences Chromosome replacement subsequence in .

[0112] (3) Use subsequences Chromosome replacement subsequences in .

[0113] Use subsequence Chromosome replacement subsequence in The replaced chromosome There may be some duplicate genes. Therefore, chromosome analysis is needed. Perform deduplication and based on chromosomes. The missing genes are filled in with the original sequence to ensure that the replaced chromosome is a valid solution.

[0114] Similarly, using subsequences Chromosome replacement subsequences in The replaced chromosome There may be some duplicate genes. Therefore, chromosome analysis is needed. Perform deduplication and based on chromosomes. The missing genes are filled in with the original sequence to ensure that the replaced chromosome is a valid solution.

[0115] After all new offspring are generated, a mutation operation is performed on the entire new population. In this mutation operation, although the probability is very low, each chromosome has a certain chance of mutating into a new chromosome. This probability is set to 1%. A chromosome mutates by randomly swapping the positions of two genes.

[0116] By implementing the three core operations of selection, crossover, and mutation in genetic algorithms, these operations work together to optimize the quality of the solution in the iterative update process of the population, ultimately approaching a high-quality task scheduling scheme.

[0117] In step 130, at least one group of heterogeneous drones is controlled to perform package loading, route planning and site access according to the task scheduling scheme to complete the delivery task.

[0118] The task sequence assigned to each drone in the high-quality task scheduling scheme is parsed into a structured task instruction that includes the order of target delivery stations, the identifier of the package to be delivered, the expected payload, and flight time constraints.

[0119] Based on the package list in the mission instructions, the system controls the automatic sorting system or manual operation terminal to load the corresponding packages into the cargo compartment of the designated drone; and monitors the total load weight in real time through the built-in weighing sensor to ensure that it does not exceed the maximum load capacity of the drone.

[0120] Based on the station visit order in the mission instructions, the current location of the UAV, electronic map data (including obstacles and no-fly zones), meteorological information (such as wind speed), and flight performance parameters (such as maximum speed and turning radius), the path planning algorithm is invoked to generate a three-dimensional safe flight path that starts from the starting point, visits each target station in sequence, and finally returns.

[0121] The generated flight path is converted into a series of waypoints and sent to the corresponding drone's flight control system via a wireless communication link (such as 4G / 5G or a dedicated remote control link). After confirming that the drone is in normal condition, the autonomous flight mode is activated to complete the delivery task.

[0122] In some implementations, the drones fly autonomously according to a waypoint sequence, performing precise landings or hovering deliveries upon arrival at each target station. After delivery, they automatically update the mission status and send a "delivered" signal and remaining battery power back to the central control system. Once all drones have completed their mission sequence and returned safely, the central control system marks the delivery mission as complete and records key indicators such as the actual maximum completion time and energy consumption for subsequent scheduling model optimization.

[0123] Please refer to Figure 3, which shows a schematic diagram of the structure of a parcel delivery device based on a drone logistics system according to an embodiment of this application. The device is applied to a drone logistics system, which includes at least one group of heterogeneous drones and multiple delivery stations. The parcel delivery device 200 based on the drone logistics system includes: a construction module 210, a generation module 220, and a delivery module 230. Specifically:

[0124] Module 210 is used to build a task scheduling model. The task scheduling model takes minimizing the maximum completion time of all package delivery tasks as the objective function, and builds a multi-dimensional set of constraints based on the heterogeneity of UAVs, flight time limits and payload capacity constraints.

[0125] The generation module 220 is used to solve the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies a set of multi-dimensional constraints.

[0126] Delivery module 230 is used to control at least one group of heterogeneous drones to perform package loading, route planning and site access according to a high-quality task scheduling scheme to complete the delivery task.

[0127] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described device and module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0128] In the several embodiments provided in this application, the coupling or direct coupling or communication connection between the modules shown or discussed may be an indirect coupling or communication connection through some interface, device or module, and may be electrical, mechanical or other forms.

[0129] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0130] Please refer to Figure 4, which shows a schematic diagram of the structure of a parcel delivery device provided in an embodiment of this application. The parcel delivery device 300 in this application may include one or more of the following components: processor 310, memory 320, and one or more application programs. The one or more application programs may be stored in memory 320 and configured to be executed by one or more processors 310. The one or more programs are configured to execute the parcel delivery method based on the UAV logistics system as described in the foregoing method embodiments.

[0131] Processor 310 may include one or more processing cores. Processor 310 connects to various parts within the entire package delivery device 300 using various interfaces and lines, and performs various functions and processes data of the package delivery device 300 by running or executing instructions, programs, code sets, or instruction sets stored in memory 320, and by calling data stored in memory 320. Optionally, processor 310 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). Processor 310 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem may also not be integrated into processor 310 and may be implemented separately through a communication chip.

[0132] The memory 320 may include random access memory (RAM) or read-only memory (ROM). The memory 320 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 320 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for implementing at least one function, instructions for implementing the various method embodiments described below, etc. The data storage area may also store data created by the package delivery device 300 during use.

[0133] Please refer to Figure 5, which shows a schematic diagram of the structure of a computer-readable storage medium provided in an embodiment of this application. The computer-readable storage medium 400 stores program code, which can be called by a processor to execute the package delivery method based on the drone logistics system described in the above method embodiment.

[0134] The computer-readable storage medium 400 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium 400 includes a non-transitory computer-readable storage medium. The computer-readable storage medium 400 has storage space for program code 410 that performs any of the method steps described above. This program code can be read from or written to one or more computer program devices. The program code 410 may be compressed, for example, in a suitable form.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A parcel delivery method based on an unmanned aerial vehicle (UAV) logistics system, characterized in that, The drone logistics system includes at least one group of heterogeneous drones and multiple delivery stations. The method includes: constructing a task scheduling model; the task scheduling model uses minimizing the maximum completion time of all package delivery tasks as the objective function, and constructs a multi-dimensional constraint set based on the heterogeneity of the drones, flight time limitations, and payload capacity constraints; the objective function is: in," "This refers to the number of the drones." "For drones" Number of flights, "For drones" In the The number of sites visited during this flight. "For drones" In the During the second flight, the visit to the The time at each delivery station, " is a binary variable, which is a wrapper Whether it was unloaded during the flight; using a genetic algorithm to solve the task scheduling model to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set; the step of using a genetic algorithm to solve the task scheduling model to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set includes: encoding the potential solution of the task scheduling problem into chromosomes; the chromosomes are arranged in a one-dimensional arrangement to represent the access priority sequence of all tasks to be delivered or destination stations; determining the priority of the drones according to the return time and load capacity of the drones, and arranging the drones according to the priority to form a drone priority queue; converting the chromosomes in the form of a two-dimensional matrix into multiple package allocation queues by row; each row in the package allocation queue corresponds to a destination station, and the order of the packages in each row represents their delivery priority at that station; in each roulette wheel selection In the strategy, a target package allocation queue is selected from the multiple package allocation queues, and the available drone with the highest current priority is selected from the drone priority queue for task allocation; the roulette wheel selection strategy is repeatedly executed to dynamically allocate tasks in the target package allocation queue to the selected drones, generating a feasible task allocation scheme; based on the feasible task allocation scheme, the complete flight path of the drone and its corresponding task completion time are determined; based on the objective function, the maximum value of the task completion time is used as the fitness value of the corresponding potential solution; during the population evolution process, selection operations are performed based on the fitness value, combined with crossover and mutation operations, to generate a new generation of population; after multiple generations of iteration, the high-quality task scheduling scheme is obtained; the at least one group of heterogeneous drones is controlled to perform package loading, path planning, and station access according to the high-quality task scheduling scheme to complete the delivery task.

2. The parcel delivery method based on an unmanned aerial vehicle (UAV) logistics system according to claim 1, characterized in that, The multi-dimensional constraint set includes at least one of the following: flight time sequence constraints, return time constraints to the originating delivery station constraints, initial time constraints, initial location constraints, package weight and drone payload capacity matching constraints, package destination and drone flight path matching constraints, single flight payload limit constraints, single flight time limit constraints, package allocation integrity constraints, and domain constraints of decision variables.

3. The parcel delivery method based on an unmanned aerial vehicle (UAV) logistics system according to claim 1, characterized in that, The method further includes: receiving periodic status reporting data from each of the UAVs, wherein the status reporting data includes at least one of the following: current geographic location coordinates, flight altitude, flight speed, remaining battery power, current payload weight, onboard package identification list, and progress information of executed tasks; and determining the flight time limit and the payload capacity constraint based on the status reporting data.

4. The parcel delivery method based on an unmanned aerial vehicle (UAV) logistics system according to claim 2, characterized in that, The flight time sequence constraint is as follows: ;in," "For drones" In the During the second flight, the visit to the The time at each delivery station, "For delivery stations" to delivery station The distance, "For drones" The flight speed; the time constraint for returning to the initial delivery station is: in," "For drones" In the The start time of the flight, i.e., the time of departure from the originating delivery station, "For drones" In the The last stop on the flight, i.e. Arrival time at each stop, "For delivery stations" To the originating delivery station The distance; the initial time constraint condition is: ;in," "For drones" In the The start time of the next flight; the initial position constraint is: The constraint condition for matching the package weight with the drone's payload capacity is as follows: ;in," "for the first" The weight of the package, "For drones" Maximum load capacity, "For all packages," There are one or more drones Its maximum load capacity meets the requirements. "For the package" weight No more than drones Maximum load capacity The constraints for matching the package destination with the drone flight path are as follows: ;in," "For the package" The final delivery location, "For drones" In the The first flight The location of each docking point; the single-flight load limit constraint is: The single flight time limitation constraint is as follows: ;in," "For drones" In the The start time of the next flight, "For drones" The maximum flight time; the package allocation integrity constraint is: The domain constraints of the decision variables are as follows: 。 5. A parcel delivery device based on an unmanned aerial vehicle (UAV) logistics system, characterized in that, The drone logistics system includes at least one group of heterogeneous drones and multiple delivery stations. The device includes: a construction module for constructing a task scheduling model; the task scheduling model uses minimizing the maximum completion time of all package delivery tasks as the objective function, and constructs a multi-dimensional constraint set based on the heterogeneity of the drones, flight time limitations, and payload capacity constraints; the objective function is: in," "This refers to the number of the drones." "For drones" Number of flights, "For drones" In the The number of sites visited during this flight. "For drones" In the During the second flight, the visit to the The time at each delivery station, " is a binary variable, which is a wrapper Whether it was unloaded during the flight; the generation module is used to solve the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set; the step of solving the task scheduling model using a genetic algorithm to generate a high-quality task scheduling scheme that satisfies the multi-dimensional constraint set includes: encoding the potential solution of the task scheduling problem into chromosomes; the chromosomes are arranged in a one-dimensional arrangement to represent the access priority sequence of all tasks to be delivered or destination stations; determining the priority of the drones according to the return time and load capacity of the drones, and arranging the drones according to the priority to form a drone priority queue; converting the chromosomes in the form of a two-dimensional matrix into multiple package allocation queues by row; each row in the package allocation queue corresponds to a destination station, and the order of the packages in each row represents their delivery priority at that station; in each roulette wheel selection In the strategy, a target package allocation queue is selected from the multiple package allocation queues, and the available drone with the highest current priority is selected from the drone priority queue for task allocation; the roulette wheel selection strategy is repeatedly executed to dynamically allocate tasks in the target package allocation queue to the selected drones, generating a feasible task allocation scheme; based on the feasible task allocation scheme, the complete flight path of the drone and its corresponding task completion time are determined; based on the objective function, the maximum value of the task completion time is used as the fitness value of the corresponding potential solution; during the population evolution process, selection operations are performed based on the fitness value, combined with crossover and mutation operations, to generate a new generation of population; after multiple generations of iteration, the high-quality task scheduling scheme is obtained; the delivery module is used to control the at least one group of heterogeneous drones to perform package loading, path planning, and station access according to the high-quality task scheduling scheme to complete the delivery task.

6. A parcel delivery device, characterized in that, include: One or more processors; Memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs being configured to perform the package delivery method based on the drone logistics system as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code, which can be called by a processor to execute the package delivery method based on the drone logistics system as described in any one of claims 1-4.

Images