Multi-unmanned vehicle cooperative path optimization method and device based on improved genetic algorithm

Abstract

The invention discloses a multi-unmanned vehicle cooperative path optimization method and device based on an improved genetic algorithm. On the basis of the established multi-unmanned-vehicle cooperative path optimization model, firstly, a k-means clustering algorithm is introduced in a population initialization stage, the search range of the understanding space is effectively reduced, a good foundation is laid for subsequent optimization, secondly, macroscopic task-vehicle matching grouping is achieved through construction of a double-layer optimization structure and a top-layer genetic algorithm, and the optimal path of the multi-unmanned-vehicle cooperative path is obtained. According to the technical scheme, refined path planning is carried out on each vehicle in the solution space of the top layer group through the bottom layer genetic algorithm, the decomposition strategy greatly reduces the search space of each time of optimization, the solving efficiency and precision are improved, and therefore the path precision and the global exploration capability are effectively improved.

Description

Technical Field

[0001] This application belongs to the field of unmanned vehicle technology, and in particular relates to a method and device for multi-unmanned vehicle cooperative path optimization based on an improved genetic algorithm. Background Technology

[0002] As a core representative of modern intelligent equipment, unmanned vehicles (RVs), with their strong carrying capacity, long endurance, and good environmental adaptability, have become key tools driving technological innovation in multiple fields such as logistics, rescue, and urban services. The complexity of their application scenarios and the diversity of task requirements are growing exponentially. For example, in post-disaster material distribution and rescue missions, RVs need to plan operating routes and collaborate with other RVs or drones to complete material delivery and personnel search and rescue; in smart city logistics distribution, multiple RVs need to efficiently cover large areas to achieve automated transportation from warehouses to last-mile stations. These scenarios pose unprecedented challenges to the autonomy, multi-vehicle collaborative capabilities, and adaptability to complex environments of RVs.

[0003] Currently, in the field of autonomous vehicle path planning, traditional optimization methods are mainly relied upon, which are difficult to meet the requirements of complex environments. Especially in scenarios such as multi-vehicle collaboration and real-time path planning, the limitations of traditional methods are becoming increasingly apparent. Specifically, these limitations manifest as unreasonable target point allocation, low overall system efficiency, and insufficient convergence performance of algorithms under complex constraints. When dealing with target optimization in multi-vehicle collaboration, traditional algorithms suffer from poor initial solution quality, slow convergence speed, and a tendency to get trapped in local optima. Summary of the Invention

[0004] This application provides a method and apparatus for multi-unmanned vehicle cooperative path optimization based on an improved genetic algorithm.

[0005] In a first aspect, embodiments of this application provide a multi-unmanned vehicle cooperative path optimization method based on an improved genetic algorithm, including: Based on the task area of ​​the autonomous vehicle and the set of target points for the tasks to be completed within the task area, a multi-autonomous vehicle cooperative path optimization model is established, wherein the set of target points includes several task nodes; the objective function of the multi-autonomous vehicle cooperative path optimization model is used to characterize the total cost of the autonomous vehicle. The multi-unmanned vehicle cooperative path optimization model is solved using a top-level genetic algorithm, and a vehicle-task matching set is output. The fitness function of the top-level genetic algorithm is the reciprocal of the objective function. The initial population of the top-level genetic algorithm is generated by clustering the task nodes according to their geographical locations using a k-means clustering algorithm. Using a low-level genetic algorithm, with the vehicle-task matching set as the initial population of the low-level genetic algorithm, the optimal path corresponding to each unmanned vehicle in the vehicle-task matching set is determined.

[0006] Optionally, each of the task nodes is defined with a corresponding task time window. ; The objective function of the multi-unmanned vehicle cooperative path optimization model is: in, This represents the total cost of the driverless car; Indicates the first weighting coefficient; This represents the second weighting coefficient; This represents the cost of the autonomous vehicle's path length. Indicates fuel consumption; This indicates the unit fuel cost; This indicates the fuel consumption per unit distance for driverless vehicles; Indicates that the driverless car starts from the node To the node driving distance, For nodes Two-dimensional plane coordinates, For nodes Two-dimensional plane coordinates; , Represents the set of target points. K represents the number of task nodes in the target point set, and K represents the set of autonomous vehicles. For variables of 0 and 1, when When, it means The vehicle completes the task node. Then proceed to the task completion node. ,when 0 indicates The vehicle completes the task node. Do not proceed to the task node. ; This indicates the penalty for delays caused by autonomous vehicles per unit of time. This indicates that the autonomous vehicle has actually reached the task node. The time.

[0007] Optionally, the constraints of the multi-unmanned vehicle cooperative path optimization model include the time constraint for an unmanned vehicle to reach any task node, the time relationship constraint for an unmanned vehicle to reach any two task nodes, and the constraint on the number of tasks completed by an unmanned vehicle in a single trip. The time constraint for the unmanned vehicle to reach any task node is: The time relationship constraint for the autonomous vehicle to reach any two task nodes is: in, This indicates that the autonomous vehicle has actually reached the task node. Time, M represents the operating speed of the autonomous vehicle, and M represents the penalty factor, which is used to constrain the time it takes for the vehicle to reach the task node. The constraint on the number of tasks that the unmanned vehicle can complete in a single trip is: in, This indicates the maximum mission capacity of an autonomous vehicle in a single operation.

[0008] Optionally, the two-dimensional planar coordinates of the task node are obtained by transforming latitude and longitude coordinates using the Mercator projection method according to the following formula: in, The transformed planar coordinates, Original latitude and longitude, For reference meridian, This is the average radius of the Earth.

[0009] Optionally, the step of using a top-level genetic algorithm to solve the multi-unmanned vehicle cooperative path optimization model and outputting a vehicle-task matching set includes: Based on the preset task requirements and the maximum task capacity of the unmanned vehicle in a single operation, determine the minimum number of clusters that meet the preset task requirements; Several task nodes are randomly selected from the target point set as initial cluster centers, and the remaining task nodes are assigned to the cluster of the cluster center that is closest to them in terms of geographical location. The cluster centers are used to represent the vehicle-task matching set. Calculate the average geographical location of all task nodes in each cluster, use it as the updated cluster center, and return to the step of assigning the remaining task nodes to the cluster center with the closest geographical location, until the preset convergence condition is met, and obtain the final vehicle-task matching set.

[0010] Optionally, the crossover operation of the top-level genetic algorithm adopts a fragment crossover method, including: randomly selecting two parent chromosomes as the target individuals for the crossover operation; randomly selecting two crossover points on the parent chromosomes to form a crossover fragment; exchanging the corresponding crossover fragments and deleting points that are repeated with the crossover fragments to form new offspring chromosomes; The mutation operation of the top-level genetic algorithm adopts the fragment mutation method, which includes: randomly selecting a chromosome as the target individual for the mutation operation; randomly selecting four mutation points on the parent chromosome to form two mutation fragments; and exchanging the positions of the two mutation fragments to form a new offspring chromosome.

[0011] Secondly, embodiments of this application provide a multi-unmanned vehicle cooperative path optimization device based on an improved genetic algorithm, comprising: A module is established to build a multi-autonomous vehicle cooperative path optimization model based on the task area of ​​the autonomous vehicle and the set of target points for the tasks to be completed within the task area. The set of target points includes several task nodes. The objective function of the multi-autonomous vehicle cooperative path optimization model is used to characterize the total cost of the autonomous vehicle. The solution module is used to solve the multi-unmanned vehicle cooperative path optimization model using a top-level genetic algorithm and output a vehicle-task matching set. The fitness function of the top-level genetic algorithm is the reciprocal of the objective function. The initial population of the top-level genetic algorithm is generated by clustering the task nodes according to their geographical locations using a k-means clustering algorithm. The determination module is used to determine the optimal path for each unmanned vehicle in the vehicle-task matching set by using the underlying genetic algorithm, with the vehicle-task matching set as the initial population of the underlying genetic algorithm.

[0012] Thirdly, embodiments of this application provide an electronic device, which includes: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, it implements the steps of the multi-vehicle cooperative path optimization method based on an improved genetic algorithm as described in any embodiment of the first aspect.

[0013] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the steps of the multi-vehicle cooperative path optimization method based on an improved genetic algorithm as described in any embodiment of the first aspect.

[0014] Fifthly, embodiments of this application provide a computer program product, which is stored in a storage medium and executed by at least one processor to implement the steps of the multi-unmanned vehicle cooperative path optimization method based on an improved genetic algorithm provided in the first aspect of embodiments of this application.

[0015] The multi-unmanned vehicle cooperative path optimization method and apparatus based on the improved genetic algorithm in this application are based on the established multi-unmanned vehicle cooperative path optimization model. First, k-means clustering algorithm is introduced in the initial population stage to effectively reduce the search range of the solution space and lay a good foundation for subsequent optimization. Second, by constructing a two-layer optimization structure, the top-level genetic algorithm realizes macro-level task-vehicle matching and grouping, and the bottom-level genetic algorithm performs fine-grained path planning for each vehicle in the solution space of the top-level group. This decomposition strategy greatly reduces the search space of each optimization, improves the solution efficiency and accuracy, and thus effectively improves the path accuracy and global exploration capability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a multi-unmanned vehicle cooperative path optimization method based on an improved genetic algorithm provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the process of solving a multi-unmanned vehicle cooperative path optimization model using a top-level genetic algorithm, as provided in an embodiment of this application. Figure 3 This is a schematic diagram of the process for generating the initial population using k-means according to an embodiment of this application; Figure 4 This is a flowchart illustrating the crossover operation of the top-level genetic algorithm provided in the embodiments of this application; Figure 5 This is a flowchart illustrating the mutation operation of the top-level genetic algorithm provided in the embodiments of this application; Figure 6 This is a flowchart illustrating the process of determining the optimal path for an autonomous vehicle using a low-level genetic algorithm, as provided in an embodiment of this application. Figure 7 This is a flowchart illustrating the mutation operation of the underlying genetic algorithm provided in the embodiments of this application; Figure 8 This is a schematic diagram of a vehicle driving path scheme obtained based on an embodiment of this application. Figure 9 These are the iterative convergence curves of the improved genetic algorithm and the traditional genetic algorithm provided in the embodiments of this application; Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0018] Figure label: Electronic device 300, processor 301, memory 302, communication interface 303, bus 310. Detailed Implementation

[0019] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0020] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0021] As a core representative of modern intelligent equipment, unmanned vehicles (RVs), with their strong carrying capacity, long endurance, and good environmental adaptability, have become key tools driving technological innovation in multiple fields such as logistics, rescue, and urban services. The complexity of their application scenarios and the diversity of task requirements are growing exponentially. For example, in post-disaster material distribution and rescue missions, RVs need to plan operating routes and collaborate with other RVs or drones to complete material delivery and personnel search and rescue; in smart city logistics distribution, multiple RVs need to efficiently cover large areas to achieve automated transportation from warehouses to last-mile stations. These scenarios pose unprecedented challenges to the autonomy, multi-vehicle collaborative capabilities, and adaptability to complex environments of RVs.

[0022] In the field of autonomous vehicle path planning, solutions can be mainly divided into two types: exact solutions and approximate solutions. Exact solutions typically utilize precise algorithms, such as the branch and bound method. This method systematically divides the solution space (branches), estimates the bounds (bounds), and discards invalid branches (pruning) to gradually approach and find the optimal solution. These algorithms are based on rigorous mathematical principles and proceed step-by-step according to deterministic rules and steps, aiming to obtain the globally optimal solution. These algorithms possess optimality and verifiability, making them suitable for scenarios with high solution quality requirements. However, as the complexity of the problem increases, the difficulty of solving it rises sharply, making it difficult to obtain a feasible solution within a finite time, thus limiting its application in large-scale problems. Approximate solutions typically employ various algorithms, such as the ant colony algorithm. This algorithm simulates the mechanism of ants using pheromones to communicate foraging paths and gradually converges to find a solution through a positive feedback mechanism. This type of algorithm obtains a feasible or near-optimal solution in a finite time with lower computational complexity. While these algorithms typically cannot guarantee global optimality, they yield high-quality solutions that meet practical needs, making them particularly suitable for large-scale vehicle path optimization problems.

[0023] Currently, task allocation and path optimization for autonomous vehicles still mainly rely on traditional optimization methods, which are insufficient to meet the requirements of complex environments. Especially in scenarios such as multi-vehicle collaboration and real-time path planning, the limitations of traditional methods are becoming increasingly apparent, specifically manifested in the following problems: Firstly, the initial solutions of the algorithm are of poor quality, resulting in slow convergence. Traditional random initialization methods (such as ant colony optimization and particle swarm optimization) generate the initial population completely randomly, lacking utilization of prior knowledge about the problem. Under complex constraints (such as time windows and load limits), this randomness leads to a large number of infeasible initial solutions. The algorithm needs to spend a large number of iterations in the early stages to eliminate inferior solutions, significantly reducing the convergence speed and making it difficult to achieve timely responses.

[0024] Secondly, the algorithms lack global optimization capabilities. Traditional algorithms derive their algorithms step by step based on deterministic rules and steps. When faced with complex environments, they suffer from problems such as computational explosion, making them difficult to apply to large-scale real-time scenarios. While swarm intelligence algorithms (such as ant colony algorithms) enhance global search capabilities through randomness, the population diversity decreases in the later stages of iteration, leading to premature convergence and difficulty in finding the optimal path in complex scenarios.

[0025] Thirdly, the solution speed is slow and the number of convergence steps is too large. The computational complexity of many optimization algorithms increases with the problem size (such as the number of task points or vehicles). The solution space that the algorithm needs to evaluate expands explosively, and the computational burden increases non-linearly and sharply. More iterations are required to converge, resulting in excessively long problem-solving time and making it difficult to find a satisfactory solution within an effective time.

[0026] To address the problems in the related technologies, this application provides a method and apparatus for multi-unmanned vehicle cooperative path optimization based on an improved genetic algorithm.

[0027] The following description, in conjunction with the accompanying drawings, details the multi-vehicle cooperative path optimization method based on an improved genetic algorithm provided in this application, through specific embodiments and application scenarios.

[0028] Figure 1 A flowchart illustrating a multi-unmanned vehicle cooperative path optimization method based on an improved genetic algorithm, according to an embodiment of this application, is shown. Figure 1 As shown, the multi-unmanned vehicle cooperative path optimization method based on the improved genetic algorithm can specifically include the following steps: S101. Based on the task area of ​​the unmanned vehicle and the set of target points for the tasks to be completed within the task area, establish a multi-unmanned vehicle cooperative path optimization model, wherein the set of target points includes several task nodes; the objective function of the multi-unmanned vehicle cooperative path optimization model is used to characterize the total cost of the unmanned vehicle. S102. Solve the multi-unmanned vehicle cooperative path optimization model using a top-level genetic algorithm and output a vehicle-task matching set. The fitness function of the top-level genetic algorithm is the reciprocal of the objective function. The initial population of the top-level genetic algorithm is generated by clustering the task nodes according to their geographical locations using a k-means clustering algorithm. S103. Using a low-level genetic algorithm, with the vehicle-task matching set as the initial population of the low-level genetic algorithm, determine the optimal path corresponding to each unmanned vehicle in the vehicle-task matching set.

[0029] Therefore, the multi-unmanned vehicle cooperative path optimization method and device based on the improved genetic algorithm, based on the established multi-unmanned vehicle cooperative path optimization model, firstly introduces the k-means clustering algorithm in the initial population stage, which effectively reduces the search range of the solution space and lays a good foundation for subsequent optimization. Secondly, by constructing a two-layer optimization structure, the top-level genetic algorithm realizes macro-level task-vehicle matching and grouping, and the bottom-level genetic algorithm performs fine-grained path planning for each vehicle in the solution space of the top-level group. This decomposition strategy greatly reduces the search space of each optimization, improves the solution efficiency and accuracy, and thus effectively improves the path accuracy and global exploration capability.

[0030] The specific implementation methods for each of the above steps are described below.

[0031] In some embodiments, in S101, the task scenario is set The task area, in which there is There are several target points (task nodes) to be completed for each task, and the set of target points is set as follows: .

[0032] Specifically, each task node has a corresponding task time window defined. When the driverless car arrives earlier than The system will incur certain penalty costs if the vehicle is later than [the specified time]. If the time limit is reached, the task cannot be completed. The autonomous vehicle's departure from the depot and completion of tasks at each coordinate is considered the completion of the entire collaborative task.

[0033] As an optional implementation, the coordinates of each task node are represented in two-dimensional coordinates, that is, the position of the node is determined by the horizontal and vertical coordinates. In practical applications, latitude and longitude coordinates are usually provided, so it is necessary to convert the latitude and longitude coordinates into two-dimensional plane coordinates before calculation. The Mercator projection method can usually be used to convert latitude and longitude, thereby ensuring the accuracy of node coordinates and the convenience of calculation. The calculation formula of Mercator projection is shown in the following formulas (1) and (2): in, The transformed planar coordinates, Original latitude and longitude, Using the reference meridian as a reference, a parking lot can be selected. This is the average radius of the Earth.

[0034] Furthermore, the driving path of the autonomous vehicle is defined as follows: , Let the sequence consist of a series of ordered task points, assuming A path planning segment is based on task points and If the path is composed of these components, the length of this path segment can be calculated as shown in equation (3): in, For nodes and The length of the road segment between them They are nodes x and y coordinates For nodes Horizontal and vertical coordinates.

[0035] Furthermore, an objective function is constructed for the multi-autonomous vehicle cooperative path optimization model to characterize the total cost of the autonomous vehicles.

[0036] First, define the path length cost of the autonomous vehicle as... Path cost is related to the actual operating distance of the autonomous vehicle and its fuel consumption. Fuel consumption The total fuel consumption of a vehicle can be calculated by multiplying the fuel consumption per unit distance by the distance traveled. The functional expression is shown in equation (4). The total transportation cost of the vehicle is the product of the total fuel consumption of all vehicles and the unit oil price. Therefore, the total transportation cost is... The function expression is shown in equation (5). in, This refers to the fuel consumption per unit distance for driverless vehicles. For nodes To the node driving distance, For unit fuel cost, For variables of 0 and 1, when ,express The vehicle completed its mission. Then proceed to complete the mission. Conversely, it means The vehicle completed its mission. Do not proceed with the mission Location, K represents the assembly of driverless vehicles.

[0037] Next, we define the time cost of the autonomous vehicle task as... Time cost refers to the penalty incurred when an autonomous vehicle completes its task later than the designated time. The time window penalty cost function for the vehicle is expressed as follows (6): in, Penalties for delays caused by driverless vehicles per unit of time. For the actual completion of the autonomous vehicle The time limit for the task.

[0038] Therefore, the total cost of the system can be derived from the above costs. The functional expression is shown in equation (7), which serves as the objective function of the multi-unmanned vehicle cooperative path optimization model: in, The total system cost, As the first weighting coefficient, This is the second weighting coefficient.

[0039] In some optional embodiments, the constraints of the multi-autonomous vehicle cooperative path optimization model include the time constraint for an autonomous vehicle to reach any task node, the time relationship constraint for an autonomous vehicle to reach any two task nodes, and the constraint on the number of tasks completed by an autonomous vehicle in a single trip. The specific definitions are as follows: Due to the time constraints of the task, the time for the unmanned vehicle to arrive at the task point must not be later than the task end time, as shown in equation (8). That is, the time constraint for the unmanned vehicle to arrive at any task node is as follows: Due to the limitation of the autonomous vehicle's own operating speed, the relationship between the arrival times of the vehicles at the two task nodes is shown in equation (9), that is, the time relationship constraint for the autonomous vehicle to arrive at any two task nodes: in, M represents the operating speed of the autonomous vehicle, and M represents the penalty factor used to constrain the time it takes for the vehicle to reach the task node.

[0040] Due to the limitations of the autonomous vehicle's own performance, there is an upper limit to the number of tasks it can complete in a single trip, which must not exceed the maximum task capacity. The expression for this is shown in equation (10), which is the constraint on the number of tasks that the autonomous vehicle can complete in a single trip: in, This represents the maximum mission capacity of a single vehicle operation.

[0041] Furthermore, in some embodiments, before S102, the parameters of the improved genetic algorithm are initialized, including: the population size, crossover rate, mutation rate, maximum number of iterations, and early termination conditions of the top-level and bottom-level genetic algorithms.

[0042] refer to Figure 2 This is a flowchart illustrating the process of solving the multi-autonomous vehicle cooperative path optimization model using a top-level genetic algorithm in S102. Figure 2 As shown, the specific steps include: First, a top-level chromosome is designed. The chromosome representing the task-matching set within this top-level chromosome is used to assign all task points to different vehicles. The chromosome length is... The chromosome sequence is The zero point represents the end point of the vehicle service, and a zero point index set is defined. Then the task set for each vehicle is .

[0043] Then, the population is initialized. Specifically, a portion of the initial population is generated using the k-means algorithm, while the remaining population is generated randomly. Figure 3 The diagram illustrates the process of generating the initial population using k-means, including the following steps: Step 3.1: Determine the minimum number of clusters To ensure that all task requirements are met, The value calculation function is expressed as shown in equation (11): in: To find the minimum number of clusters required to meet all task requirements, This represents the maximum mission capacity of a single vehicle operation.

[0044] Step 3.2: Randomly select from all tasks One task serves as the initial cluster center, with each center representing a vehicle-task matching set. The remaining tasks will be assigned to the cluster center closest to them in the next step.

[0045] Step 3.3: Calculate the distance of each task to all cluster centers and assign the task to the set containing the nearest cluster center to ensure that the geographical locations of the task points in each set are as close as possible.

[0046] Step 3.4: Recalculate the center point of each set, that is, use the average position of all tasks in the set as the new center position for the redistribution of tasks in the next round.

[0047] Step 3.5: Set a convergence criterion, such as the center point no longer changing, that is, after several consecutive iterations, the position of the cluster center is basically stable or the maximum number of iterations is reached. If the set convergence criterion is met, the calculation is terminated and the final task-vehicle matching set is generated; otherwise, repeat Step 3.3 until the grouping structure is stable.

[0048] In other words, based on the preset task requirements and the maximum task capacity of the autonomous vehicle in a single operation, the minimum number of clusters required to meet the preset task requirements is determined; several task nodes are randomly selected from the target point set as initial cluster centers, and the remaining task nodes are assigned to the cluster containing the cluster center with the closest geographical location. The cluster center is used to represent the vehicle-task matching set; the average geographical location of all task nodes in each cluster is calculated and used as the updated cluster center, and the process returns to the step of assigning the remaining task nodes to the cluster containing the cluster center with the closest geographical location, until the preset convergence condition is met, and the final vehicle-task matching set is obtained.

[0049] In this way, the k-means spatial clustering algorithm is used to generate a portion of the high-quality initial population in the top-level genetic algorithm. This method pre-groups the data based on the geographical distribution characteristics of the task points, making the initial solution closer to the optimal solution and avoiding the blindness caused by completely random initialization, thus laying a solid foundation for the algorithm's rapid convergence.

[0050] Back Figure 2 In the top-level optimization process, fitness evaluation is performed. Specifically, the fitness function is the reciprocal of the total system cost (i.e., ), and the goal is to minimize the total system cost.

[0051] Furthermore, the selection operation employs an elite retention strategy, preserving individuals with higher fitness to ensure that high-quality solutions can be passed on to the next generation.

[0052] Furthermore, the crossover operation employs a segmented crossover method. Specifically, Figure 4 The flowchart of the crossover operation in the top-level genetic algorithm is shown, including the following steps: Step 4.1: Randomly select two parent chromosomes as the target individuals for the crossover operation.

[0053] Step 4.2: Randomly select two crossover sites on the parent chromosome to form a crossover segment.

[0054] Step 4.3: Exchange the corresponding crossover segments and delete the points that are repeated with the crossover segments to form new offspring chromosomes.

[0055] Furthermore, the mutation operation employs fragment mutation. Specifically, Figure 5 The flowchart of the mutation operation in the top-level genetic algorithm is shown, including the following steps: Step 5.1: Randomly select a chromosome as the target individual for the mutation operation.

[0056] Step 5.2: Randomly select four mutation sites on the parent chromosome to form two mutation segments. Step 5.3: Swap the positions of these two segments to form a new offspring chromosome.

[0057] Back Figure 2 Finally, after multiple iterations, the top-level genetic optimization operation ends when the set termination condition is met, ultimately obtaining the driving path and departure time for each autonomous vehicle. Specifically, the underlying genetic algorithm is used to optimize the vehicle-task matching set output by the top-level genetic algorithm, ultimately obtaining the actual running path of each vehicle. Substituting the path into the model, the departure time can be calculated.

[0058] refer to Figure 6 This is a flowchart illustrating the process of determining the optimal path for the autonomous vehicle using a low-level genetic algorithm in S103. Figure 6 As shown, the specific steps include: First, chromosomes are designed. The solution for bottom-level path optimization consists of the optimal path for a vehicle to complete a series of task points. Since the number of tasks completed by each vehicle at the top level is different, the length of the chromosome is uncertain. Optionally, in this embodiment, the inner-level path optimization chromosomes adopt a symbolic encoding method based on natural numbers, and each chromosome represents the running path of a vehicle.

[0059] Then, the vehicle-task matching set generated by the top-level optimization is used as input as the initial population of the bottom layer.

[0060] Furthermore, during the optimization of the underlying path, a fitness evaluation is performed. Specifically, the fitness function is the reciprocal of the total cost of the sub-path.

[0061] Furthermore, the selection operation uses a roulette wheel selection method; the crossover operation uses a segment crossover method, and... Figure 4 The process shown is the same; the mutation operation uses the improved cycle mutation method, specifically... Figure 7 The flowchart illustrating the mutation operation of the underlying genetic algorithm is shown, including the following steps: Step 7.1: Randomly select a parent chromosome as the target individual for the mutation operation.

[0062] Step 7.2: Randomly select a mutation point and swap it with other points in turn. Calculate the fitness of the individual after the swap and determine if it has improved. If the fitness has improved, proceed to step 3; otherwise, continue swapping with the next point until a better solution is found or all points are traversed.

[0063] Step 7.3: Exchange sites to form new chromosomes.

[0064] Back Figure 6 Finally, when the number of iterations reaches the preset termination condition, the inner-layer genetic optimization operation ends, and the optimal solution of the set is obtained, which is the actual running route of each vehicle.

[0065] In other words, this embodiment constructs a two-layer optimization structure. A top-level genetic algorithm achieves macroscopic task-vehicle matching and grouping, while a bottom-level genetic algorithm performs refined path planning for each vehicle within the solution space of the top-level grouping. Specifically, this includes: defining model and algorithm parameters, setting algorithm parameters including population size, crossover rate, mutation rate, maximum number of iterations, and early termination conditions to ensure the stability of the model solution and the optimization capability of the algorithm; using the k-means algorithm to generate the initial solution for the top-level genetic algorithm, while simultaneously using a random generation method to generate other initial solutions to ensure population diversity and enhance the algorithm's global search capability; using the vehicle-task matching set generated by the outer optimization layer as input to the inner genetic algorithm for further path optimization; randomly shuffling the received vehicle-task matching set to generate the initial solution for the inner genetic algorithm; evaluating the fitness of the current path to reflect the quality of the solution; and using an elite retention strategy based on individual fitness to retain some high-quality solutions for the next generation to ensure the transmission of excellent genes in the population and the convergence speed of the algorithm. Crossover and mutation operations form the next generation population (i.e., the new vehicle task completion order); record the current iteration number and increment the iteration number by one; determine if the termination condition of the underlying genetic algorithm is met; if the set termination condition is met, terminate; otherwise, continue optimization; after the underlying path planning is completed, determine the optimal service order and recalculate the overall fitness to provide an optimization basis for the top-level genetic algorithm; according to the individual fitness, adopt an elite retention strategy to retain some high-quality solutions for the next generation; perform crossover and mutation operations to generate the next generation population of the top-level algorithm (i.e., the new vehicle-task matching set) to further optimize the overall scheme; increase the iteration number of the top-level genetic algorithm to continuously optimize the vehicle-task matching scheme; determine if the termination condition of the top-level genetic algorithm is met; if the set termination condition is met, terminate and output the optimal vehicle driving scheme; otherwise, continue optimization.

[0066] Furthermore, to verify the effectiveness of the multi-autonomous vehicle cooperative path optimization method based on the improved genetic algorithm in this application embodiment, experimental analysis was conducted using the C104 case from the Solomon standard test dataset, which is widely used in the field of vehicle path planning. Experimental results show that the optimal solution obtained by the multi-autonomous vehicle cooperative path optimization method based on the improved genetic algorithm in this application embodiment is 903.48 yuan, while the optimal solution obtained by the traditional genetic algorithm is 1651.71 yuan. Figure 8 The present invention demonstrates a vehicle travel route scheme obtained based on the embodiments of this application, showing reasonable route planning and scheduling efficiency. Figure 9 The convergence curve comparison shows that the multi-unmanned vehicle cooperative path optimization method based on the improved genetic algorithm in this application embodiment exhibits excellent performance in the early stage of iteration and has a faster convergence speed. This proves that the embodiment of this application has strong global optimization capability and solution efficiency in large-scale complex problems.

[0067] Experimental results show that the embodiments of this application significantly improve the solution quality compared with traditional genetic algorithms, increasing the optimal solution rate by 45.3%, fully verifying the effectiveness of the improved strategy of this application. It can be understood that the improved algorithm performance is mainly due to the k-means initialization strategy, the two-layer optimization structure, and the problem-oriented genetic operation design. These improvements effectively enhance the algorithm's global exploration capability and convergence performance.

[0068] The proposed method in this application achieves significant improvements in both overall performance and application scenarios compared to traditional ant colony optimization, particle swarm optimization, and the original genetic algorithm. First, k-means clustering is introduced in the initialization phase for initial population optimization, effectively avoiding the low solution quality and slow convergence problems caused by random initialization in traditional algorithms, thus laying a high-quality starting point for subsequent optimization. Second, the algorithm adopts a hierarchical collaborative architecture: at the top level, an efficient but imprecise algorithm is used to quickly allocate tasks; at the bottom level, each vehicle performs local fine-tuning based on the top-level results and real-time environmental information. This mechanism significantly enhances the algorithm's global optimization capability in complex dynamic scenarios. Furthermore, a modified cycle mutation operation is introduced during the bottom-level optimization process, further accelerating the convergence process and reducing unnecessary search overhead, enabling the algorithm to significantly improve convergence speed while maintaining solution quality.

[0069] In summary, the embodiments of this application employ two processes: task allocation and path planning. The outer layer optimizes the task-vehicle matching set, while the inner layer optimizes the specific path planning. The algorithm introduces the k-means clustering method for initial population optimization, effectively avoiding the problems of low initial solution quality and slow convergence in traditional genetic algorithms. This enhances its ability to collaboratively process tasks and optimizes various practical constraints (such as endurance, loading limitations, and task time requirements) and cost objectives (such as path length, time cost, and energy consumption) to improve the overall system operating efficiency and reduce operating costs. This has important theoretical and practical significance for promoting the development of multi-unmanned vehicle collaboration, providing a new solution for unmanned vehicle operation, and also helping to improve solution efficiency and enhance the applicability of the algorithm in practical applications.

[0070] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0071] Based on the same technical concept, corresponding to any of the above embodiments, this application also provides a multi-unmanned vehicle cooperative path optimization device based on an improved genetic algorithm.

[0072] Specifically, the multi-unmanned vehicle cooperative path optimization device based on the improved genetic algorithm may include: A module is established to build a multi-autonomous vehicle cooperative path optimization model based on the task area of ​​the autonomous vehicle and the set of target points for the tasks to be completed within the task area. The set of target points includes several task nodes. The objective function of the multi-autonomous vehicle cooperative path optimization model is used to characterize the total cost of the autonomous vehicle. The solution module is used to solve the multi-unmanned vehicle cooperative path optimization model using a top-level genetic algorithm and output a vehicle-task matching set. The fitness function of the top-level genetic algorithm is the reciprocal of the objective function. The initial population of the top-level genetic algorithm is generated by clustering the task nodes according to their geographical locations using a k-means clustering algorithm. The determination module is used to determine the optimal path for each unmanned vehicle in the vehicle-task matching set by using the underlying genetic algorithm, with the vehicle-task matching set as the initial population of the underlying genetic algorithm.

[0073] It should be noted that, for ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this application, the functions of each module can be implemented in one or more software and / or hardware.

[0074] The apparatus in the above embodiments is used to implement the corresponding multi-vehicle cooperative path optimization method based on the improved genetic algorithm in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0075] Based on the same technical concept, corresponding to any of the above embodiments, this application also provides an electronic device.

[0076] Figure 10 A schematic diagram of a more specific electronic device hardware structure provided in this embodiment is shown.

[0077] The electronic device 300 may include a processor 301 and a memory 302 storing computer program instructions.

[0078] Specifically, the processor 301 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0079] Memory 302 may include mass storage for data or instructions. For example, and not limitingly, memory 302 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 302 may include removable or non-removable (or fixed) media. Where appropriate, memory 302 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 302 is non-volatile solid-state memory.

[0080] In certain embodiments, the memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Thus, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of this application.

[0081] The processor 301 reads and executes computer program instructions stored in the memory 302 to implement any of the multi-vehicle cooperative path optimization methods based on the improved genetic algorithm in the above embodiments.

[0082] In some examples, the electronic device 300 may also include a communication interface 303 and a bus 310. For example, Figure 10 As shown, the processor 301, memory 302, and communication interface 303 are connected through bus 310 and complete communication with each other.

[0083] The communication interface 303 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0084] Bus 310 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not as a limitation, bus 310 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 310 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0085] For example, the electronic device 300 can be a mobile phone, tablet computer, laptop computer, handheld computer, in-vehicle electronic device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc.

[0086] Based on the same technical concept, corresponding to any of the methods in the above embodiments, this application also provides a non-transitory computer-readable storage medium. This computer-readable storage medium stores computer program instructions; when executed by a processor, these computer program instructions implement any of the multi-vehicle cooperative path optimization methods based on an improved genetic algorithm in the above embodiments. Examples of computer-readable storage media include non-transitory computer-readable storage media, such as portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, etc.

[0087] Based on the same technical concept, corresponding to any of the above embodiments, this application also provides a computer program product, which includes computer program instructions. In some embodiments, the computer program instructions can be executed by one or more processors of a computer to cause the computer and / or the processors to execute the multi-unmanned vehicle cooperative path optimization method based on the improved genetic algorithm. Corresponding to the execution entity for each step in each embodiment of the multi-unmanned vehicle cooperative path optimization method based on the improved genetic algorithm, the processor executing the corresponding step can belong to the corresponding execution entity.

[0088] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0089] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0090] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0091] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0092] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for cooperative path optimization of multiple unmanned vehicles based on an improved genetic algorithm, characterized in that, include: Based on the task area of ​​the autonomous vehicle and the set of target points for the tasks to be completed within the task area, a multi-autonomous vehicle cooperative path optimization model is established, wherein the set of target points includes several task nodes; the objective function of the multi-autonomous vehicle cooperative path optimization model is used to characterize the total cost of the autonomous vehicle. The multi-unmanned vehicle cooperative path optimization model is solved using a top-level genetic algorithm, and a vehicle-task matching set is output. The fitness function of the top-level genetic algorithm is the reciprocal of the objective function. The initial population of the top-level genetic algorithm is generated by clustering the task nodes according to their geographical locations using a k-means clustering algorithm. Using a low-level genetic algorithm, with the vehicle-task matching set as the initial population of the low-level genetic algorithm, the optimal path corresponding to each unmanned vehicle in the vehicle-task matching set is determined.

2. The method according to claim 1, characterized in that, Each of the aforementioned task nodes has a corresponding task time window defined. ; The objective function of the multi-unmanned vehicle cooperative path optimization model is: in, This represents the total cost of the driverless car; Indicates the first weighting coefficient; This represents the second weighting coefficient; This represents the cost of the autonomous vehicle's path length. Indicates fuel consumption; This indicates the unit fuel cost; This indicates the fuel consumption per unit distance for driverless vehicles; Indicates that the driverless car starts from the node To the node driving distance, For nodes Two-dimensional plane coordinates, For nodes Two-dimensional plane coordinates; , Represents the set of target points. K represents the number of task nodes in the target point set, and K represents the set of autonomous vehicles. For variables of 0 and 1, when When, it means Vehicle completes task node Then proceed to the task completion node. ,when 0 indicates Vehicle completes task node Do not proceed to the task node. ; This indicates the penalty for delays caused by autonomous vehicles per unit of time. This indicates that the autonomous vehicle has actually reached the task node. The time.

3. The method according to claim 2, characterized in that, The constraints of the multi-unmanned vehicle cooperative path optimization model include the time constraint for an unmanned vehicle to reach any task node, the time relationship constraint for an unmanned vehicle to reach any two task nodes, and the constraint on the number of tasks completed by an unmanned vehicle in a single trip. The time constraint for the unmanned vehicle to reach any task node is: The time relationship constraint for the autonomous vehicle to reach any two task nodes is: in, This indicates that the autonomous vehicle has actually reached the task node. Time, M represents the operating speed of the autonomous vehicle, and M represents the penalty factor, which is used to constrain the time it takes for the vehicle to reach the task node. The constraint on the number of tasks that the unmanned vehicle can complete in a single trip is: in, This indicates the maximum mission capacity of an autonomous vehicle in a single operation.

4. The method according to claim 2, characterized in that, The two-dimensional planar coordinates of the task node are obtained by transforming latitude and longitude coordinates using the Mercator projection method according to the following formula: in, The transformed planar coordinates, Original latitude and longitude, For reference meridian, This is the average radius of the Earth.

5. The method according to claim 1, characterized in that, The method utilizes a top-level genetic algorithm to solve the multi-autonomous vehicle cooperative path optimization model, outputting a vehicle-task matching set, including: Based on the preset task requirements and the maximum task capacity of the unmanned vehicle in a single operation, determine the minimum number of clusters that meet the preset task requirements; Several task nodes are randomly selected from the target point set as initial cluster centers, and the remaining task nodes are assigned to the cluster of the cluster center that is closest to them in terms of geographical location. The cluster centers are used to represent the vehicle-task matching set. Calculate the average geographical location of all task nodes in each cluster, use it as the updated cluster center, and return to the step of assigning the remaining task nodes to the cluster center with the closest geographical location, until the preset convergence condition is met, and obtain the final vehicle-task matching set.

6. The method according to claim 1, characterized in that, The crossover operation of the top-level genetic algorithm adopts a segment crossover method, which includes: randomly selecting two parent chromosomes as the target individuals for the crossover operation; randomly selecting two crossover points on the parent chromosomes to form a crossover segment; exchanging the corresponding crossover segments and deleting points that are repeated with the crossover segments to form new offspring chromosomes; The mutation operation of the top-level genetic algorithm adopts the fragment mutation method, which includes: randomly selecting a chromosome as the target individual for the mutation operation; randomly selecting four mutation points on the parent chromosome to form two mutation fragments; and exchanging the positions of the two mutation fragments to form a new offspring chromosome.

7. A multi-unmanned vehicle cooperative path optimization device based on an improved genetic algorithm, characterized in that, include: A module is established to build a multi-autonomous vehicle cooperative path optimization model based on the task area of ​​the autonomous vehicle and the set of target points for the tasks to be completed within the task area. The set of target points includes several task nodes. The objective function of the multi-autonomous vehicle cooperative path optimization model is used to characterize the total cost of the autonomous vehicle. The solution module is used to solve the multi-unmanned vehicle cooperative path optimization model using a top-level genetic algorithm and output a vehicle-task matching set. The fitness function of the top-level genetic algorithm is the reciprocal of the objective function. The initial population of the top-level genetic algorithm is generated by clustering the task nodes according to their geographical locations using a k-means clustering algorithm. The determination module is used to determine the optimal path for each unmanned vehicle in the vehicle-task matching set by using the underlying genetic algorithm, with the vehicle-task matching set as the initial population of the underlying genetic algorithm.

8. An electronic device, characterized in that, The device includes: a processor and a memory storing computer program instructions; when the processor invokes the computer program instructions, it implements the multi-unmanned vehicle cooperative path optimization method based on an improved genetic algorithm as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when invoked by a processor, implement the multi-unmanned vehicle cooperative path optimization method based on an improved genetic algorithm as described in any one of claims 1-6.

10. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device performs the multi-unmanned vehicle cooperative path optimization method based on the improved genetic algorithm as described in any one of claims 1-6.

Images