Multi-unmanned aerial vehicle task allocation method and system based on multi-target blue whale hunting mechanism

Abstract

The invention provides a multi-unmanned aerial vehicle task allocation method and system based on a multi-target blue whale hunting mechanism, and belongs to the field of unmanned aerial vehicle task allocation. The problems that an existing unmanned aerial vehicle multi-target task allocation method is prone to falling into local optimum, and later convergence precision is insufficient are solved. Segmented initialization is applied during initialization, initialization is performed according to task allocation in the first stage, so that an initial blue whale group is more reasonable, and a chaotic mapping idea is adopted in the second stage, so that initial distribution of the blue whale group has higher randomness and ergodicity; an elitism strategy is applied, excellent individuals of each generation are reserved, and the convergence speed is increased; a simulated annealing thought is introduced in each iteration process, so that a relatively high mutation probability exists in the initial stage of iteration to accept a relatively poor solution so as to jump out of local optimum; the blue whale positions are searched in a continuous space through continuous value coding, threshold conversion and an optimization algorithm, a blue whale group is guided to evolve to an excellent solution through fitness function feedback, and finally an actually executable task allocation scheme is obtained through mapping.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) task allocation technology, and more specifically, to a multi-UAV task allocation method and system based on a multi-target blue whale hunting mechanism. Background Technology

[0002] In recent years, drone technology has developed rapidly and has been widely used in disaster relief, material delivery, and environmental monitoring. As a crucial component of drone systems, efficient, safe, and collaborative task allocation is a pressing technical challenge that urgently needs to be addressed. Drone task allocation refers to assigning mission objectives to drones under various complex environmental conditions and constraints to meet diverse mission requirements.

[0003] Currently, the main methods for solving the task allocation problem include: traditional optimization algorithms, heuristic optimization algorithms, multi-agent systems, and reinforcement learning. Heuristic optimization algorithms have significant advantages over other methods in solving multi-objective optimization problems. Furthermore, drone task allocation also faces numerous technical challenges: how to adapt to dynamic environments, how to handle large-scale drone collaboration, how to reduce communication latency, and how to prevent malicious interference.

[0004] In their paper "A Review of Research on Multi-UAV Cooperative Planning" published in Computer Engineering and Applications (2025, 61(1): 1002-8331), Ning Cong et al. pointed out that multi-objective optimization problems involve trade-offs among multiple conflicting objectives, such as shortest path versus minimum energy consumption, minimum task time versus maximum task benefit, etc. The research focus is on how to find a balanced solution or Pareto optimal solution among multiple objectives to meet different task requirements and constraints. In their 2017 paper "A Multi-Objective Optimization Algorithm for Solving EngineerDesign Problems", Pradeep Jangir and Narottam Jangir proposed a new multi-objective whale optimization algorithm (Non-Dominated Sorting Whale Optimization Algorithm, NSWOA) based on the original whale optimization algorithm (WOA), which can solve multi-objective problems, providing a new method for solving multi-objective problems. This method uses few parameters, converges quickly, and is easy to implement. However, due to the inefficient file update mechanism in the multi-objective whale optimization algorithm, if the external files are occupied by inferior solutions or locally non-dominated solutions in the early stages, the randomly selected leader will spread this inferior information to the entire population, forming a vicious cycle. This causes the algorithm to perform fine-grained searches in the wrong areas, making the method prone to insufficient convergence accuracy in the later stages. Secondly, the inherent defects of the original whale optimization algorithm are amplified in the multi-objective whale optimization algorithm. For example, the spiral update simulates the bubble attack of whales, which makes individuals move in a spiral between the leader and their current position. This mechanism is very effective when exploring the area around the leader, but its long-distance exploration ability is relatively weak. When the entire population is confined to a certain local area, it is difficult to escape the local convergence problem by relying solely on spiral update. Summary of the Invention

[0005] The technical problem to be solved by this invention is:

[0006] To address the problem that existing UAV multi-target task allocation methods are prone to getting trapped in local optima and have insufficient convergence accuracy in the later stages.

[0007] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0008] This invention provides a multi-UAV task allocation method based on a multi-target blue whale hunting mechanism, comprising the following steps:

[0009] S100. Establish a multi-UAV task allocation model, including constructing two objective functions and three constraints. The two objective functions are used to maximize the UAV task completion rate and minimize the total rescue time of the UAV, respectively. When constructing the constraints, a penalty term is introduced, which includes a penalty for the UAV exceeding the maximum load, a penalty for the target not being rescued or exceeding the number of rescue attempts, and a penalty for the UAV exceeding the maximum endurance.

[0010] S200: The Blue Whale swarm is initialized using a segmented initialization strategy, including the first stage where each task is assigned at least one qualified drone, and the second stage where the task is randomly initialized and assigned using the idea of ​​chaotic mapping.

[0011] S300. Calculate the fitness function value of an individual blue whale, including obtaining the blue whale's location coordinates and mapping them to a task assignment matrix, and calculating the fitness function value of the corresponding blue whale from the task assignment matrix.

[0012] S400. Calculate and sort the non-dominant rank based on the fitness function values ​​of all blue whales;

[0013] S500: Calculate the crowding distance of blue whales of the same Pareto class and fill the Pareto front archive until the archive is full. Then select the first blue whale from the archive as the best individual.

[0014] S600. Update the position of the blue whale pod through the group evolution mechanism of the blue whale pod, wherein the group evolution mechanism includes the spiral mechanism, the search and foraging mechanism and the shrinking and encircling mechanism;

[0015] S700: Based on the elite retention strategy, the historical high-quality Blue Whale engine is retained;

[0016] S800, applying simulated annealing criteria, updates the position of the Blue Whale again;

[0017] S900. Check if the preset iteration limit has been reached. If it has been reached, end the iteration and convert the location information of the blue whale with the highest non-dominance level into a task allocation matrix to obtain the final task allocation scheme. If it has not been reached, return to step S600 to continue execution until the iteration limit is reached.

[0018] Further, in step S100,

[0019] It has A drone that needs to perform a mission, using U n Representing the 100 drones, the target number is ,use It means that T m Representing the The task allocation matrix between the drone and the target is as follows: ,in , , Indicates the first Drone rescue One goal, Indicates the first The drone is not the first Rescue efforts will be carried out on each target;

[0020] After introducing a penalty term, the multi-drone task allocation model becomes: ,

[0021] in,

[0022] ;

[0023] ;

[0024] in, Prioritizing the rescue mission Indicates the rescue Resources required for each objective For the first The maximum number of relief supplies that a drone can carry. For the first The total flight range of the drone during the mission Indicates the first The average moving speed of the drone For the first The maximum flight time of the drone Let be the fitness function of the established model. Let the task completion fitness function be used. Let the fitness function be the total rescue time. It is a function with maximum value. These are the weighting factors for the UAV payload constraint penalty, the target rescue constraint penalty, and the UAV range constraint penalty, respectively.

[0025] Further, in step S200,

[0026] Set the number of individual blue whales to The maximum number of iterations is The branch coefficients of the Logistic mapping are defined as follows: ; will the first The generation The location of the blue whale is recorded as follows ,in The chaotic mapping formula is ,in, For the first The generation The first blue whale Dimensional position, Randomly generated within the defined domain. , ;

[0027] For ease of calculation, the first... The generation The fitness vector of a blue whale is denoted as ,in The current iteration number is taken initially. , For the first The generation Only the blue whale Fitness values.

[0028] Further, in step S300,

[0029] Location of the blue whale Measurement is The measurement relationship is ,in , , To determine the threshold, The blue whale's position on the row vector is the first... The position coordinates of the dimension are mapped to the task allocation matrix. Each blue whale represents a task allocation scheme, determined by the task allocation matrix. The generation The fitness function of a blue whale is ,in .

[0030] Furthermore, in step S400, the method for calculating the non-dominant level is as follows:

[0031] Let the first The two blue whales from that era, numbered respectively and When the blue whale Neither of their fitness levels is comparable to that of the blue whale. When the time is wrong, then the blue whale Dominating the Blue Whale This process involves traversing the entire blue whale population and recording its non-dominance level: individuals not dominated by any blue whale are placed in the first level with a non-dominance level of 1; blue whales dominated only by blue whales in the first level are placed in the second level; and so on, until all blue whales have been divided, and finally arranged in ascending order of non-dominance level.

[0032] Further, in step S500, the method for calculating the congestion distance is as follows:

[0033] In the first level, the blue whales in the first level are sorted in ascending order according to the first fitness value. After sorting, the crowding distance between the two boundaries is set to infinity. Then, the distance between adjacent individuals is calculated, which is the difference in fitness value. The fitness values ​​of the two boundaries are used to normalize it. Finally, the second fitness value is traversed, and the two normalized values ​​are added together to obtain the crowding distance.

[0034] After obtaining the crowding distance of each blue whale, individuals in the same Pareto class are sorted in descending order of crowding distance; finally, the blue whales in the first class are stored in the Pareto front archive in turn until the archive is full, and then the first blue whale in the archive is selected as the best individual.

[0035] Further, in step S600,

[0036] The blue whale pod will use the best individual blue whale obtained in step S500 as its target. The optimal position of the blue whale is denoted as First, generate a range within Random numbers between And update control number The coefficient of surrounding prey ;

[0037] when and At that moment, the blue whale pod contracted to surround their prey. The generation The formula for updating the position of a blue whale is: ,in , The number that decreases linearly from 2 to 0 with the number of iterations. , and for Random numbers between For the number of iterations, The upper limit of iterations;

[0038] when and At that time, the blue whale performs a spiral motion, and the position update formula is: ,in Helical factor, Used to simulate spiral motion. , for Random numbers between;

[0039] when At that time, the blue whale pod randomly searches, and the position update formula is: , It is in the Blue whales were randomly selected from the blue whale pod.

[0040] Further, in step S700, the elite retention strategy is to merge the two generations of blue whale pods into a single pod of size [missing information]. After performing fitness calculations, non-dominated ranking, and crowding ranking on the group, the top... Only the elite blue whales were retained to form a new blue whale pod.

[0041] Further, in step S800, the simulated annealing criterion is that the blue whale pod accepts a solution with a poor historical record with a jump probability, thereby escaping the local optimum. The jump probability is... ,in The current temperature. The initial temperature, The temperature decay coefficient is , For the first The generation The first blue whale fitness, The label for the Blue Whale that performed poorly in the previous generation was randomly selected from the differential solutions.

[0042] A multi-UAV task allocation system based on a multi-target blue whale hunting mechanism has program modules corresponding to the above steps, and executes the steps in the multi-UAV task allocation method based on the multi-target blue whale hunting mechanism when running.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] This invention employs a segmented initialization approach during initialization. The first stage initializes the blue whale swarm according to task allocation, resulting in a more reasonable initial swarm. The second stage uses chaotic mapping to enhance the randomness and ergodicity of the initial distribution. An elite retention strategy is applied to preserve the best individuals in each generation, increasing the convergence speed. Simulated annealing is introduced in each iteration, allowing for a higher mutation probability in the early stages to accept poor solutions and escape local optima. As the number of iterations increases, the mutation probability decreases, and local search is performed in the later stages to increase accuracy. The blue whale positions are indirectly represented by continuous value encoding and threshold conversion to represent the task allocation scheme. The optimization algorithm searches in a continuous space, guiding the blue whale swarm towards better solutions through fitness function feedback. Finally, a practically executable task allocation scheme is obtained through mapping, balancing the algorithm's exploration capability with the problem's discreteness requirements. Attached Figure Description

[0045] Figure 1 This is a flowchart of a multi-UAV task allocation method based on a multi-target blue whale hunting mechanism in an embodiment of the present invention;

[0046] Figure 2The result of the UAV task allocation scheme obtained using this method in the embodiments of the present invention. Figure 1 ;

[0047] Figure 3 The result of the UAV task allocation scheme obtained by this method in the embodiments of the present invention. Figure 2 ;

[0048] Figure 4 The result of the UAV task allocation scheme obtained using this method in the embodiments of the present invention. Figure 3 . Detailed Implementation

[0049] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0050] Specific Implementation Plan 1: Combining Figure 1 As shown, this invention provides a multi-UAV task allocation method based on a multi-target blue whale hunting mechanism, comprising the following steps:

[0051] S100, Establish a multi-UAV task allocation model

[0052] It has A drone that needs to perform a mission, using U n Representing the 100 drones, the target number is ,use It means that T m Representing the The task allocation matrix between the drone and the target is as follows: ,in , , Indicates the first Drone rescue One goal, Indicates the first The drone is not the first Rescue efforts will be carried out on each target;

[0053] The multi-UAV task allocation model constructed in this invention has two objective functions and three constraints. To simplify the calculation, a penalty term is introduced, and the constraints are transformed into penalty terms and substituted into the objective functions. The optimization objectives are: to maximize the UAV task completion rate and minimize the total rescue time of the UAVs. The penalty terms are: penalty for UAVs exceeding the maximum load, penalty for targets not being rescued or exceeding the number of rescue attempts, and penalty for UAVs exceeding the maximum endurance.

[0054] The mission objectives and the drones should also adhere to the following constraints:

[0055] 1. Mission constraints: Each mission can only be completed by one drone;

[0056] 2. Constraints on drones: Each drone may carry only a limited amount of relief supplies; drones may not exceed their maximum range;

[0057] The following is a simplified model for multi-drone task allocation: ,

[0058] in, , ,in Prioritizing the rescue mission Indicates the rescue Resources required for each objective For the first The maximum number of relief supplies that a drone can carry. For the first The total flight range of the drone during the mission Indicates the first The average moving speed of the drone For the first The maximum flight time of the drone Let be the fitness function of the established model. Let the task completion fitness function be used. Let the fitness function be the total rescue time. It is a function with maximum value. , and These are the weighting factors for the UAV payload constraint penalty, the target rescue constraint penalty, and the UAV range constraint penalty, respectively.

[0059] S200, Initialize the Blue Whale Swarm

[0060] To generate a more reasonable initial blue whale swarm, a segmented initialization strategy is adopted. The first stage of initialization is used to ensure that each task is assigned at least one qualified drone. The second stage is used to ensure population diversity. Based on the first stage, the chaotic mapping idea is used to randomly initialize and assign tasks.

[0061] Set the number of individual blue whales to The maximum number of iterations is The branch coefficients of the Logistic mapping are defined as follows: ; will the first The generation The location of the blue whale is recorded as follows ,in The chaotic mapping formula is ,in, For the first The generation The first blue whale Dimensional position, Randomly generated within the defined domain. , For ease of calculation, the first... The generation The fitness vector of a blue whale is denoted as ,in The current iteration number is taken initially. , For the first The generation Only the blue whale Each fitness value;

[0062] S300, Calculate the fitness function value of an individual blue whale.

[0063] The calculation process is as follows: [The blue whale's location is...] Measurement is The measurement relationship is ,in , , To determine the threshold, a value of 0.5 is used. The blue whale's position on the row vector is the first... The position coordinates of the dimension are mapped to the task allocation matrix. Each blue whale represents a task allocation scheme, determined by the task allocation matrix. The generation The fitness function of a blue whale is ,in ;

[0064] S400. Calculate the non-dominant rank based on the fitness function values ​​of all blue whales, and then sort them.

[0065] The non-dominant hierarchy is calculated as follows: (The rest of the text appears to be a list of numbers and symbols, possibly related to a calculation or system.) The two blue whales from that era, numbered respectively and When the blue whale Neither of their fitness levels is comparable to that of the blue whale. When the time is poor, it is called a blue whale. Dominating the Blue Whale This process involves traversing the entire blue whale group and recording its non-dominance level: individuals not dominated by any blue whale are placed in the first level with a non-dominance level of 1; blue whales dominated only by blue whales in the first level and not by any other individuals are placed in the second level; and so on, until all blue whales have been divided, and finally arranged in ascending order of non-dominance level.

[0066] S500, calculate the crowding distance of blue whales of the same Pareto optimal level and populate the Pareto front archive.

[0067] The crowding distance is calculated as follows: Taking the blue whales in the first level as an example, first, the blue whales in the first level are sorted in ascending order according to the first fitness value. After sorting, the crowding distance between the two boundaries is set to infinity. Then, the distance between adjacent individuals is calculated, which is the difference in fitness values. The fitness values ​​of the two boundaries are then normalized. Finally, the second fitness value is traversed, and the two normalized values ​​are added together to obtain the desired crowding distance. After obtaining the crowding distance of each blue whale, individuals in the same Pareto level are sorted in descending order according to the crowding distance. Finally, the blue whales in the first level are stored sequentially in the Pareto front archive until the archive is full. Then, the first blue whale in the archive is selected as the optimal individual.

[0068] S600 updates the location of the blue whale pod through the group evolution mechanism of the blue whale pod.

[0069] There are three main mechanisms for the evolution of blue whale pods: the spiral mechanism, the search and foraging mechanism, and the shrinking and encircling mechanism. Blue whale pods evolve according to these three mechanisms with a certain probability.

[0070] The blue whale pod will use the best individual blue whale obtained in step S500 as its target. The optimal position of the blue whale is denoted as First, generate a range within Random numbers between And update control number The coefficient of surrounding prey ;when and At that moment, the blue whale pod contracted to surround their prey. The generation The formula for updating the position of a blue whale is: ,in , The number that decreases linearly from 2 to 0 with the number of iterations. , and for Random numbers between For the number of iterations, The upper limit of iterations; when and At that time, the blue whale performs a spiral motion, and the position update formula is: ,in, Helical factor, Used to simulate spiral motion. , for Random numbers between; when At that time, the blue whale pod randomly searches, and the position update formula is: , It is in the Blue whales randomly selected from a pod of blue whales;

[0071] S700, Elite Retention Strategy: Retaining Historically High-Quality Blue Whale

[0072] Elite Preservation: Merging two generations of blue whale pods into a single entity of size... After performing fitness calculations, non-dominated ranking, and crowding ranking on the group, the top... Only the elite blue whales are retained to form a new blue whale pod;

[0073] S800, applying simulated annealing criteria, further updates the location of the Blue Whale.

[0074] Simulated annealing acceptance criterion: The blue whale pod accepts a poorly historical solution with a certain jump probability, thus escaping the local optimum. The jump probability is... ,in The current temperature. The initial temperature, The temperature decay coefficient is , For the first The generation The first blue whale A fitness value, The labels for the Blue Whales that performed poorly in the previous generation were randomly selected from the differential solutions;

[0075] S900, Check if the preset iteration limit has been reached. If the condition has been met, the iteration ends and proceeds to the next step; if the condition has not been met, then... Return to step six and continue execution; convert the location information of the blue whale with the highest non-dominance level into a task allocation matrix to obtain the final task allocation scheme.

[0076] Specific Implementation Scheme 2: The present invention provides a multi-UAV task allocation system based on a multi-target blue whale hunting mechanism. The system has program modules corresponding to the above steps, and executes the steps in the above-mentioned multi-UAV task allocation method based on a multi-target blue whale hunting mechanism when running.

[0077] The other combinations and connections in this implementation scheme are the same as in Specific Implementation Scheme 1.

[0078] Simulation Experiment

[0079] In the simulation experiment, Figures 2 to 4 To obtain the Pareto optimal solution using this method, combined with Figures 2 to 4 As shown, set the group size. Maximum number of iterations Simulated annealing initial temperature Temperature decay coefficient Branch coefficients of the Logistic mapping The penalty coefficients related to load limits, mission limits, and range limits are respectively taken as follows: This invention uses a 0-1 programming model to solve the multi-UAV task allocation problem, setting the total number of UAVs. The total number of targets that need to be attacked The drone starts from the origin, and the target's two-dimensional coordinates are... The maximum payload of the drone is The maximum range of the drone is The resource requirements for mission rescue are The priority of each task is as follows: The average speed of the drone is The parameters are as follows:

[0080]

[0081]

[0082] The multi-target UAV task allocation scheme obtained using this method is shown in Table 1. In Table 1... Represents the drone's serial number. The objective function 1 represents the mission number, and the objective functions 1 and 2 represent the drone mission completion rate and the total rescue time, respectively. Objective function 1 is the minimum value optimization objective obtained by transforming it through the reciprocal operation.

[0083] Table 1

[0084]

[0085] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A multi-UAV task allocation method based on a multi-target blue whale hunting mechanism, characterized in that, Includes the following steps: S100. Establish a multi-UAV task allocation model, including constructing two objective functions and three constraints. The two objective functions are used to maximize the UAV task completion rate and minimize the total rescue time of the UAV, respectively. When constructing the constraints, a penalty term is introduced, which includes a penalty for the UAV exceeding the maximum load, a penalty for the target not being rescued or exceeding the number of rescue attempts, and a penalty for the UAV exceeding the maximum endurance. S200: The Blue Whale swarm is initialized using a segmented initialization strategy, including the first stage where each task is assigned at least one qualified drone, and the second stage where the task is randomly initialized and assigned using the idea of ​​chaotic mapping. S300. Calculate the fitness function value of an individual blue whale, including obtaining the blue whale's location coordinates and mapping them to a task assignment matrix, and calculating the fitness function value of the corresponding blue whale from the task assignment matrix. S400. Calculate and sort the non-dominant rank based on the fitness function values ​​of all blue whales; S500: Calculate the crowding distance of blue whales of the same Pareto class and fill the Pareto front archive until the archive is full. Then select the first blue whale from the archive as the best individual. S600. Update the position of the blue whale pod through the group evolution mechanism of the blue whale pod, wherein the group evolution mechanism includes the spiral mechanism, the search and foraging mechanism and the shrinking and encircling mechanism; S700: Based on the elite retention strategy, the historical high-quality Blue Whale engine is retained; S800, applying simulated annealing criteria, updates the position of the Blue Whale again; S900. Check if the preset iteration limit has been reached. If it has been reached, end the iteration and convert the location information of the blue whale with the highest non-dominance level into a task allocation matrix to obtain the final task allocation scheme. If the limit is not reached, return to step S600 and continue execution until the iteration limit is reached.

2. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 1, characterized in that: In step S100, It has A drone that needs to perform a mission, using U n Representing the 100 drones, the target number is ,use It means that T m Representing the The task allocation matrix between the drone and the target is as follows: ,in , , Indicates the first Drone rescue One goal, Indicates the first The drone is not the first Rescue efforts will be carried out on each target; After introducing a penalty term, the multi-drone task allocation model becomes: , in, ； ； in, Prioritizing the rescue mission Indicates the rescue Resources required for each objective For the first The maximum number of relief supplies that a drone can carry. For the first The total flight range of the drone during the mission Indicates the first The average moving speed of the drone For the first The maximum flight time of the drone Let be the fitness function of the established model. Let the task completion fitness function be used. Let the fitness function be the total rescue time. It is a function with maximum value. These are the weighting factors for the UAV payload constraint penalty, the target rescue constraint penalty, and the UAV range constraint penalty, respectively.

3. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 2, characterized in that: In step S200, Set the number of individual blue whales to The maximum number of iterations is The branch coefficients of the Logistic mapping are defined as follows: ; will the first The generation The location of the blue whale is recorded as follows ,in The chaotic mapping formula is ,in, For the first The generation The first blue whale Dimensional position, Randomly generated within the defined domain. , ; For ease of calculation, the first... The generation The fitness vector of a blue whale is denoted as ,in The current iteration number is taken initially. , For the first The generation Only the blue whale Fitness values.

4. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 3, characterized in that: In step S300, Location of the blue whale Measurement is The measurement relationship is ,in , , To determine the threshold, The blue whale's position on the row vector is the first... The position coordinates of the dimension are mapped to the task allocation matrix. Each blue whale represents a task allocation scheme, determined by the task allocation matrix. The generation The fitness function of a blue whale is ,in .

5. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 4, characterized in that: In step S400, the non-dominant level is calculated as follows: Let the first The two blue whales from that era, numbered respectively and When the blue whale Neither of their fitness levels is comparable to that of the blue whale. When the time is wrong, then the blue whale Dominating the Blue Whale This process involves traversing the entire blue whale population and recording its non-dominance level: individuals not dominated by any blue whale are placed in the first level with a non-dominance level of 1; blue whales dominated only by blue whales in the first level are placed in the second level; and so on, until all blue whales have been divided, and finally arranged in ascending order of non-dominance level.

6. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 5, characterized in that: In step S500, the method for calculating the congestion distance is as follows: In the first level, the blue whales in the first level are sorted in ascending order according to the first fitness value. After sorting, the crowding distance between the two boundaries is set to infinity. Then, the distance between adjacent individuals is calculated, which is the difference in fitness value. The fitness values ​​of the two boundaries are used to normalize it. Finally, the second fitness value is traversed, and the two normalized values ​​are added together to obtain the crowding distance. After obtaining the crowding distance of each blue whale, individuals in the same Pareto class are sorted in descending order of crowding distance; finally, the blue whales in the first class are stored in the Pareto front archive in turn until the archive is full, and then the first blue whale in the archive is selected as the best individual.

7. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 6, characterized in that: In step S600, The blue whale pod will use the best individual blue whale obtained in step S500 as its target. The optimal position of the blue whale is denoted as First, generate a range within Random numbers between And update control number The coefficient of surrounding prey ; when and At that moment, the blue whale pod contracted to surround their prey. The generation The formula for updating the position of a blue whale is: ,in , The number that decreases linearly from 2 to 0 with the number of iterations. , and for Random numbers between For the number of iterations, The upper limit of iterations; when and At that time, the blue whale performs a spiral motion, and the position update formula is: ,in Helical factor, Used to simulate spiral motion. , for Random numbers between; when At that time, the blue whale pod randomly searches, and the position update formula is: , It is in the Blue whales were randomly selected from the blue whale pod.

8. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 7, characterized in that: In step S700, the elite retention strategy is to merge two generations of blue whale pods into a single pod of size [missing information]. After performing fitness calculations, non-dominated ranking, and crowding ranking on the group, the top... Only the elite blue whales were retained to form a new blue whale pod.

9. The multi-UAV task allocation method based on the multi-target blue whale hunting mechanism according to claim 8, characterized in that: In step S800, the simulated annealing criterion is that the blue whale pod accepts a solution with a poor historical record with a jump probability, thereby escaping the local optimum. The jump probability is... ,in The current temperature. The initial temperature, The temperature decay coefficient is , For the first The generation The first blue whale fitness, The label for the Blue Whale that performed poorly in the previous generation was randomly selected from the differential solutions.

10. A multi-UAV task allocation system based on a multi-target blue whale hunting mechanism, characterized in that: The system has a program module corresponding to the steps of any one of the claims 1-9 above, and executes the steps in the above-described multi-UAV task allocation method based on the multi-target blue whale hunting mechanism when it is run.

Images