A method for unmanned aerial vehicle cluster task allocation based on improved multi-objective ant colony algorithm

Abstract

The invention discloses a method for allocating tasks of a drone cluster based on an improved multi-objective ant colony algorithm, comprising: constructing two ant groups and two corresponding heuristic information respectively according to an optimization target one and an optimization target two; dividing each ant group into a plurality of ant subgroups to construct an ant cluster; each drone competes for the execution right of the next task based on the pheromone of its members; the drone that obtains the execution right of the next task selects the next task according to a state transition rule; updating the local pheromone only when the ant cluster completes constructing a task allocation solution; updating the global pheromones of the two ant groups only when all the ant clusters complete constructing the task allocation solution; when the iteration is completed, all solutions in the archive set constitute the final task allocation plan, and the drone cluster executes the task according to the task allocation plan.

Description

Technical Field

[0001] The present invention belongs to the technical field of task allocation, and specifically relates to a method for allocating tasks of a drone cluster based on an improved multi-objective ant colony algorithm. Background Art

[0002] With the rapid development of drone technology, drone swarms, as an efficient and flexible system, have received extensive attention and research. When performing tasks, drone swarms can make full use of their flexibility and scalability to improve the efficiency and success rate of task execution through collaborative execution. The task allocation of drone swarms is a multi-objective and multi-constrained optimization problem, and its complexity increases sharply with the increase in the number of drones and the scale of tasks. When performing tasks, it is necessary to comprehensively consider factors such as the task environment, resource constraints, and multi-machine collaboration, and reasonably allocate tasks to each drone to ensure that the drone swarm can effectively perform tasks while meeting various constraints. Therefore, studying the task allocation method of drone swarms is of great significance to improving the performance of drone swarms in performing tasks.

[0003] At present, some research results have been achieved in the research on the task allocation of drone clusters, but there are still some shortcomings. First, most of the models are modeled as single-objective optimization models, or the multi-objective optimization model is converted into a single-objective optimization model through the weighted method, which makes it difficult to truly describe the complexity of task allocation when the drone cluster performs tasks. Secondly, when calculating the range cost in task allocation, only the two-dimensional Euclidean distance from the drone to the target point is considered, and the three-dimensional flight distance of the drone in the actual environment is not considered. Finally, it is not considered that when the drone cluster performs tasks, the performance characteristics of the drones, such as flight speed and endurance, are different. The above shortcomings can easily lead to a large deviation between the actual task execution effect and the expected effect when the drone executes the task allocation plan. Therefore, in the task allocation of drone clusters, the performance characteristics of each drone and the actual situation of the task environment must be fully considered to ensure the effectiveness of task allocation and maximize the task execution effect.

[0004] The ant colony algorithm is an optimization algorithm based on ant foraging behavior proposed by Dorigo. It uses a positive feedback mechanism and has strong robustness. It has shown excellent performance in solving combinatorial optimization problems. In recent years, with the increasing complexity and diversity of optimization problems, researchers have begun to pay attention to the multi-objective ant colony algorithm and proposed a variety of improved and variant algorithms to improve the convergence, diversity and robustness of the algorithm. However, the current multi-objective ant colony algorithm pays more attention to theoretical research and is difficult to be directly applied to the task allocation of drone clusters. Therefore, it is necessary to improve it so that the multi-objective ant colony algorithm can be effectively applied to the problem of drone cluster task allocation, and use the algorithm's heuristic search strategy and positive feedback mechanism to continuously adjust and optimize the quality of the solution during the solution process, so as to gradually approach the Pareto optimal solution set. Summary of the invention

[0005] Purpose of the invention: In order to solve the problems of over-simplification of the model, inaccurate calculation of the flight cost, and failure to consider the performance differences of UAVs in the existing task allocation methods, and to solve the problem that the current multi-objective ant colony algorithm is difficult to directly apply to the task allocation of UAV clusters, the present invention proposes a UAV cluster task allocation method based on an improved multi-objective ant colony algorithm.

[0006] Technical solution: A method for allocating UAV cluster tasks based on an improved multi-objective ant colony algorithm, comprising the following steps:

[0007] Step 1: The first optimization goal is to minimize the total flight distance of the drone cluster after the task is completed, and the second optimization goal is to minimize the task completion time; the total flight distance of the drone cluster after the task is completed refers to the flight distance of all drones to the task estimated based on the initial position of the drone, the location of the task and the flight environment; the task completion time refers to the completion time of the last task;

[0008] Step 2: Determine the coordination constraints between the drones in the drone cluster, determine the timing constraints for executing the tasks, and determine the flight range constraints for each drone;

[0009] Step 3: According to optimization goal 1 and optimization goal 2, two ant groups and two corresponding heuristic information are constructed respectively, and each ant group is composed of multiple ants;

[0010] Step 4: Divide each ant group into multiple ant subgroups. For any ant subgroup, each ant in it represents the same drone.

[0011] Step 5: Take one ant from each ant sub-group to form an ant cluster;

[0012] Step 6: Update the member pheromone of each drone based on its initial value and remaining flight range; each drone competes for the right to execute the next task based on its member pheromone;

[0013] Step 7: The drone that obtains the execution right of the next task selects the next task according to the state transition rule; the state transition rule is determined by the pheromone and the heuristic information in the task environment;

[0014] Step 8: Repeat steps 6 to 7 until all ants in the ant cluster have been traversed and the task allocation solution corresponding to the ant cluster is obtained; then execute step 9;

[0015] Step 9: Calculate the value of optimization objective 1 and optimization objective 2 of the task allocation solution corresponding to the ant cluster, and update the local pheromone;

[0016] Step 10: Repeat steps 5 to 9 until all ant clusters are traversed and the task allocation solutions corresponding to all ant clusters are obtained, and then proceed to step 11;

[0017] Step 11: Store the Pareto optimal solutions of all currently constructed task allocation solutions into the archive set. The currently constructed task allocation solutions refer to all solutions in the archive set of the previous iteration and the task allocation solutions corresponding to all ant clusters obtained in this iteration; find multiple Pareto optimal solutions closest to the origin in the archive set, recorded as the solution set Origin; update the global pheromones of the two ant groups according to all solutions in the archive set and all solutions in the solution set Origin;

[0018] Step 12: Determine whether the termination condition is met. If so, output all solutions in the archive set. All solutions in the archive set constitute the final task allocation plan. If not, execute step 3.

[0019] Step 13: The drone cluster executes the task according to the task allocation plan obtained in step 12.

[0020] Furthermore, the total flight distance of the drone cluster after the mission is completed is modeled according to the following steps:

[0021] Digital simulation of the mission environment using digital elevation models;

[0022] Total flight distance of the drone cluster after the mission is completed Modeled as:

[0023] ;

[0024] in, For drones On mission The flight range is estimated by using the three-dimensional range cost calculation method based on the vertical section; is the task decision variable, representing the drone Whether to execute the task , Indicates drone Not performing tasks , Indicates drone Execute the task ; represents the total number of drones, Indicates the total number of tasks;

[0025] The time to complete the task Modeled as:

[0026] ;

[0027] in, The time it takes for the drone to perform the mission, For drones flight speed.

[0028] Furthermore, in step 2, the coordination constraints between the drones in the drone cluster are expressed as:

[0029] ;

[0030] In the formula, is the task decision variable, representing the drone Whether to execute the task , Indicates drone Not performing tasks , Indicates drone Execute the task ;

[0031] This coordination constraint means that a UAV can only perform one task at a time, and a task can only be performed by one UAV;

[0032] For any two adjacent tasks that the drone is to perform With the task , the drone must complete the mission After that, you can execute the task ,Therefore, the timing constraint of executing the task is expressed as:

[0033] ;

[0034] in, , The UAV arrival mission ,Task time, and tasks The execution order of tasks Before, The time spent on the drone mission;

[0035] The flight range constraint of each UAV is expressed as:

[0036] ;

[0037] in, For drones The maximum flight range; For drones On mission The flight range is estimated by using the three-dimensional range cost calculation method based on the vertical section.

[0038] Furthermore, the two heuristic information in step 3 include the total flight distance of the drone cluster after the mission is completed. Corresponding heuristic information and task completion time Corresponding heuristic information , respectively expressed as:

[0039] ;

[0040] in, For the task Transfer to Task the cost of the voyage required; is the average flight time of all drones; For drones Current flight time; For drones From the task Transfer to Task The time cost, For drones flight speed.

[0041] Furthermore, in step 6, the member pheromone of each drone is updated according to the initial value and remaining flight range of each drone, which is expressed as:

[0042] Drone exist Member pheromones of the moment , by drone The initial value of and the remaining flight distance at this time are updated as:

[0043] ;

[0044] In the formula, For drones exist The flight distance flown at the time; For drones The initial value of For drones The maximum flight range.

[0045] Furthermore, in step 6, each drone competes for the right to execute the next task based on its member pheromones, which is expressed as:

[0046] According to member pheromones , drones exist The probability of obtaining the task execution right of the next task at any time is:

[0047] ;

[0048] In the formula, Indicates the total number of drones.

[0049] Furthermore, in step 7, the drone that obtains the execution right of the next task selects the next task according to the state transition rule, which is expressed as:

[0050] Assume that the drone that obtains the execution right of the next mission is drone , the state transition rule is expressed as:

[0051] ;

[0052] in, is a random variable generated according to the probability distribution given by:

[0053] ;

[0054] in, Ant groups, each of which has its own pheromone matrix and heuristic information; for A collection of tasks to be selected at any time; for Time Mission With the task pheromone concentration between; for Time Mission With the task The inspirational information between them; and They are pheromone factor and heuristic information factor, respectively, which are used to control the relative importance of pheromone and heuristic information; for A random number between is the balance factor, when When using the greedy method, Roulette is used.

[0055] Furthermore, in step 9, the updating of local pheromones is expressed as:

[0056] ;

[0057] in, for Time Mission With the task The pheromone concentration between is the local pheromone volatility coefficient, is the initial value of pheromone.

[0058] Furthermore, in step 11, the global pheromone of the two ant groups is updated, which is expressed as:

[0059] ;

[0060] in, The value of is non-negative; is the global pheromone volatility coefficient; After each iteration, the task With the task The pheromone increment between is calculated as:

[0061] ;

[0062] In the formula, Pass the path for all solutions in the archive The pheromones left behind; The solution path in the solution set Origin The pheromones left behind; is the pheromone constant; Origin is the solution set Origin; and They are the objective function values ​​of the solutions in the archive set and the objective function values ​​of the solutions in the solution set Origin respectively.

[0063] Furthermore, in step 10, after obtaining the task allocation solutions corresponding to all ant clusters, the state transition rules Adaptive update is performed according to the following formula:

[0064] ;

[0065] in, for The upper limit of is the population diversity set based on Shannon entropy; For path The probability of appearing in the task allocation solution; is the population size; For path The number of times it appears in the task allocation solution.

[0066] Beneficial effects: The present invention establishes a task allocation model with two optimization objectives and multiple constraints, and proposes a competition-based multi-objective ant colony system algorithm to solve the model. Compared with the prior art, the present invention has the following advantages:

[0067] (1) The method of the present invention takes into account the complexity of task allocation in task scenarios, sets two optimization objectives, and sets different heuristic information and pheromone matrices according to the two optimization objectives, providing decision makers with more options, ensuring that task allocation can be flexibly adjusted according to different situations, better adapting to changes in the task environment, and improving task allocation efficiency;

[0068] (2) The method of the present invention establishes constraints for the performance differences of UAVs and the coordination of multiple UAVs, and sets a member pheromone matrix in the improved ant colony algorithm, ensuring that only UAVs with remaining flight range can continue to perform subsequent tasks, thereby effectively satisfying the constraints in the model and making UAVs with more remaining flight range have a greater chance of being selected to perform tasks, effectively balancing the remaining flight range of UAVs and avoiding the situation where a UAV cannot continue to participate in subsequent tasks due to premature exhaustion of its remaining flight range;

[0069] (3) The method of the present invention designs adaptive transfer rules according to population diversity, so that the algorithm can accelerate convergence when the population diversity is large; when the algorithm tends to the local optimum, the search space is expanded to increase the possibility of finding a better solution;

[0070] (4) The method of the present invention sets a limited-capacity archive set to store the Pareto optimal solution, realizes information sharing among populations, and designs a pheromone update method based on the origin, thereby maintaining population diversity while improving the convergence of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 This is a flow chart of a method for allocating tasks of drone clusters based on an improved multi-objective ant colony algorithm proposed by the present invention;

[0072] Figure 2 This is a model diagram of the mountain mission environment and the initial and target positions of the UAV;

[0073] Figure 3 The figure is a schematic diagram of one of the task allocation schemes obtained by using the UAV cluster task allocation method based on the improved multi-objective ant colony algorithm proposed in the present invention. DETAILED DESCRIPTION

[0074] The technical solution of the present invention is now further described in conjunction with the accompanying drawings and embodiments.

[0075] like Figure 1 As shown, this embodiment discloses a method for allocating tasks of drone clusters based on an improved multi-objective ant colony algorithm, which mainly includes the following steps:

[0076] Step 1: Minimizing the total flight distance of the drone cluster after the mission is completed and minimizing the mission completion time are set as two optimization objectives.

[0077] Among them, the total flight range of the drone cluster after the mission is completed is the estimated flight range of all drones to the mission based on the initial position of the drone, the location of the mission and the mission environment.

[0078] In this embodiment, a digital elevation model is used to digitally simulate the mission environment, and the mission environment and the initial position of the UAV and the mission position are as follows: Figure 2 shown.

[0079] Total flight distance of the drone cluster after the mission is completed Modeled as:

[0080] ;

[0081] in, For drones On mission The flight range is estimated by using the three-dimensional range cost calculation method based on the vertical section; is the task decision variable, representing the drone Whether to execute the task , Indicates drone Not performing tasks , Indicates drone Execute the task ; represents the total number of drones, Indicates the total number of tasks;

[0082] The task completion time It refers to the completion time of the last task, which consists of two parts: UAV Time spent traveling vs. time spent performing tasks , so it is modeled as:

[0083] ;

[0084] in, For drones flight speed.

[0085] Step 2: According to the coordination between drones, the order of task execution and the remaining resources of drones, set the constraints for the drone cluster when executing the task.

[0086] A drone can only perform one task at a time, and a task can only be performed by one drone. The coordination constraints between drones in the drone cluster are modeled as follows:

[0087] ;

[0088] In the formula, is the task decision variable, representing the drone Whether to execute the task , Indicates drone Not performing tasks , Indicates drone Execute the task ;

[0089] This coordination constraint means that a UAV can only perform one task at a time, and a task can only be performed by one UAV;

[0090] For any two adjacent tasks that the drone is to perform With the task , the drone must complete the mission After that, you can execute the task ,Therefore, the timing constraint of executing the task is expressed as:

[0091] ;

[0092] in, , The UAV arrival mission ,Task time, and tasks The execution order of tasks Before, The time spent on the drone mission;

[0093] Affected by the energy consumption of the drone, the total flight range of each drone should be less than the maximum flight range of the drone, so the maximum range constraint is modeled as:

[0094] ;

[0095] in, For drones The maximum flight range.

[0096] Step 3: Initialize the parameters of the improved multi-objective ant colony algorithm, and construct two ant groups and corresponding heuristic information according to the two optimization goals. Each ant group consists of multiple ants. The specific operations include:

[0097] Initialize the parameters of the improved multi-objective ant colony algorithm, including: population size , maximum number of iterations , pheromone factors , Heuristic Information Factor , global pheromone volatility coefficient , Local pheromone volatility coefficient wait.

[0098] In order to improve the quality and search efficiency of Pareto solutions, two heuristic information are designed for the two optimization objectives. and The specific expression is:

[0099] ;

[0100] in, For the task Transfer to Task the cost of the voyage required; is the average flight time of all drones; For drones Current flight time; For drones From the task Transfer to Task The time cost, For drones To prevent the large number from eating the small number, the heuristic information is normalized.

[0101] Step 4: Divide each ant group into multiple ant subgroups. For any ant subgroup, each ant it contains represents the same drone.

[0102] Step 5: Each ant sub-group takes out an ant to form an ant cluster, which is actually a drone cluster, to form a task allocation plan. In this embodiment, the ant sub-groups and ant clusters are constructed at the same time.

[0103] Step 6: Set member pheromones for each drone based on its initial value and remaining flight range; store the member pheromones of each drone by designing a member pheromone matrix. Each drone competes for the right to execute the next mission based on the member pheromones.

[0104] Drone exist Member pheromones of the moment , by drone The initial value of and the remaining flight distance at this time are updated as:

[0105] ;

[0106] In the formula, For drones exist The flight distance flown at the time; For drones The initial value of For drones The maximum flight range.

[0107] According to member pheromones , drones exist The probability of obtaining the task execution right of the next task at any time is:

[0108] .

[0109] By setting member pheromones, this embodiment can effectively balance the value of each drone and the remaining flight range, ensure that drones with exhausted resources are no longer selected during the task allocation process, meet the constraints of task allocation, thereby reducing the generation of infeasible solutions and improving the efficiency of task allocation.

[0110] Step 7: The drone that obtains the task execution right for the next task selects the next task according to the state transition rule.

[0111] The state transition rule is determined by the pheromone distribution and heuristic information in the environment. Assume that the drone that obtains the execution right of the next task is the drone , the state transition rule is expressed as:

[0112] ;

[0113] in, is a random variable generated according to the probability distribution given by:

[0114] ;

[0115] in, Ant groups, each of which has its own pheromone matrix and heuristic information; for A collection of tasks to be selected at any time; for Time Mission With the task pheromone concentration between; for Time Mission With the task The inspirational information between them; and They are pheromone factor and heuristic information factor, respectively, which are used to control the relative importance of pheromone and heuristic information; for A random number between is the balance factor, when When using the greedy method, Roulette method is used.

[0116] Step 8: Determine whether the ant cluster has completed constructing a task allocation solution. If completed, execute step 9; if not completed, execute step 6;

[0117] Step 9: Calculate the two optimization target values ​​of this task allocation solution and update the local pheromone. The update method is:

[0118] The updating local pheromone is expressed as:

[0119] ;

[0120] in, for Time Mission With the task The pheromone concentration between is the local pheromone volatility coefficient, is the initial value of pheromone.

[0121] Step 10: Determine whether all ant clusters have completed the construction of task allocation solutions. If so, update the balance factor according to the population diversity, and then execute step 11; if not, execute step 5.

[0122] In the state transition rules, The size of largely determines the choice of the next task. Improve it so that it has adaptive properties and changes according to the diversity of the population. The calculation method is:

[0123] ;

[0124] in, for The upper limit of , to avoid the algorithm falling into the local optimum; is the population diversity set based on Shannon entropy; For path The probability of appearing in the task allocation solution; is the population size; For path The number of times it appears in the task allocation solution.

[0125] This embodiment makes the balance factor have adaptive characteristics, so that the multi-objective ant colony algorithm can be used when the population diversity is large. Increase, thereby accelerating the convergence speed; when the multi-objective ant colony algorithm tends to the local optimum, Reduce to expand the search space and increase the possibility of finding a better solution.

[0126] Step 11: Find the Pareto optimal solution among all the currently constructed task allocation solutions, and store the Pareto optimal solution among all the currently constructed task allocation solutions into the archive set. All the currently constructed task allocation solutions refer to all the solutions in the archive set of the previous iteration and all the task allocation solutions obtained in this iteration. If the number of solutions in the archive set Archive exceeds the maximum capacity of the archive set Archive, calculate the congestion degree of each solution in the archive set Archive, and remove solutions with small congestion degrees until the number of solutions in the archive set Archive is reduced to the maximum capacity of the archive set Archive.

[0127] In order to prevent each ant group from being too close to the extreme point of its own target value, some solutions closest to the origin are found in the archive set Archive, which is recorded as the solution set Origin; according to all the solutions in Archive and all the solutions in the solution set Origin, the global pheromones of the two ant groups are updated, so as to adjust and optimize the pheromones, and improve the convergence of the algorithm while maintaining diversity.

[0128] The update of global pheromone can be expressed as:

[0129] ;

[0130] in, The value of is non-negative; is the global pheromone volatility coefficient; After each iteration, the task With the task The pheromone increment between is calculated as:

[0131] ;

[0132] In the formula, Pass the path for all solutions in the archive The pheromones left behind; The solution path in the solution set Origin The pheromones left behind; is the pheromone constant; Origin is the solution set Origin; and They are the objective function values ​​of the solutions in the archive set and the objective function values ​​of the solutions in the solution set Origin. The smaller the objective function value of the solution, the more pheromones are left.

[0133] Step 12: Determine whether the termination condition is met. If so, output all solutions in the archive set Archive. All solutions in the archive set Archive constitute the final task allocation plan. If not, execute step 3.

[0134] Step 13: The drone cluster executes the task according to the task allocation plan.

[0135] In this embodiment, the UAV cluster task allocation generates multiple non-dominated solutions, and one of the task allocation solutions is as follows: Figure 3 As shown in the figure, the circle is the initial position of the UAV, the asterisk is the mission position, and the dotted line is the estimated flight track. The UAV performs the mission in the order of the track. Figure 3 It can be seen that this embodiment takes into account the three-dimensional flight distance of the UAV, and simultaneously takes into account the total flight range and mission completion time of the UAV.

Claims

1. A method for allocating tasks of drone clusters based on an improved multi-objective ant colony algorithm, characterized by: The following steps are involved: Step 1: The first optimization goal is to minimize the total flight distance of the drone cluster after the task is completed, and the second optimization goal is to minimize the task completion time; the total flight distance of the drone cluster after the task is completed refers to the flight distance of all drones to the task estimated based on the initial position of the drone, the location of the task and the flight environment; the task completion time refers to the completion time of the last task; Step 2: Determine the coordination constraints between the drones in the drone cluster, determine the timing constraints for executing the tasks, and determine the flight range constraints for each drone; Step 3: According to optimization goal 1 and optimization goal 2, two ant groups and two corresponding heuristic information are constructed respectively, and each ant group is composed of multiple ants; Step 4: Divide each ant group into multiple ant subgroups. For any ant subgroup, each ant in it represents the same drone. Step 5: Take one ant from each ant sub-group to form an ant cluster; Step 6: Update the member pheromone of each drone based on its initial value and remaining flight range; each drone competes for the right to execute the next task based on its member pheromone; Step 7: The drone that obtains the execution right of the next task selects the next task according to the state transition rule; the state transition rule is determined by the pheromone and the heuristic information in the task environment; Step 8: Repeat steps 6 to 7 until all ants in the ant cluster have been traversed and the task allocation solution corresponding to the ant cluster is obtained; Then proceed to step 9; Step 9: Calculate the value of optimization objective 1 and optimization objective 2 of the task allocation solution corresponding to the ant cluster, and update the local pheromone; Step 10: Repeat steps 5 to 9 until all ant clusters are traversed and the task allocation solutions corresponding to all ant clusters are obtained, and then proceed to step 11; Step 11: Store the Pareto optimal solutions of all currently constructed task allocation solutions into the archive set. The currently constructed task allocation solutions refer to all solutions in the archive set of the previous iteration and the task allocation solutions corresponding to all ant clusters obtained in this iteration; find multiple Pareto optimal solutions closest to the origin in the archive set, recorded as the solution set Origin; update the global pheromones of the two ant groups according to all solutions in the archive set and all solutions in the solution set Origin; Step 12: Determine whether the termination condition is met. If so, output all solutions in the archive set. All solutions in the archive set constitute the final task allocation plan. If not satisfied, go to step 3; Step 13: The drone cluster executes the task according to the task allocation plan obtained in step 12; The two heuristic information in step 3 include the total flight distance F of the drone cluster after the mission is completed path Corresponding heuristic information and task completion time F time Corresponding heuristic information Respectively expressed as: Among them, l lj For the task T l Transfer to task T j the cost of the voyage required; is the average flight time of all UAVs; t i For UAV i Current flight time; For UAV i From Task T l Transfer to task T j The time cost, For UAV i Flight speed; In step 6, the member pheromone of each drone is updated according to the initial value and remaining flight range of each drone, which is expressed as: UAV i Member pheromone γ at time t i (t), by UAV U i The initial value of and the remaining flight distance at this time are updated as: In the formula, For UAV i The distance flown at time t; For UAV i The initial value of For UAV i The maximum flight range; In step 7, the drone that obtains the execution right of the next task selects the next task according to the state transition rule, which is expressed as: Assume that the drone that obtains the execution right for the next mission is drone U i , the state transition rule is expressed as: Where J is a random variable generated according to the probability distribution given by the following formula: Among them, * is an ant group, each ant group has its own pheromone matrix and inspiration information; allowed t is the set of tasks to be selected at time t; is the task T at time t l With task T j pheromone concentration between; is the task T at time t l With task T j heuristic information between; α and β are pheromone factor and heuristic information factor respectively, which are used to control the relative importance of pheromone and heuristic information; q is a random number between [0,1]; q0∈[0,1] is the balance factor, when q≤q0, the greedy method is used, and when q>q0, the roulette method is used.

2. The method for allocating unmanned aerial vehicle cluster tasks based on an improved multi-objective ant colony algorithm according to claim 1 is characterized in that: The total flight distance of the drone cluster after the mission is completed is modeled according to the following steps: Digital simulation of the mission environment using digital elevation models; The total flight distance of the drone cluster after the mission is completed F path Modeled as: Among them, d ij For UAV i In the execution of task T j The flight range is estimated by using the three-dimensional range cost calculation method based on the vertical section; X ij is the task decision variable, representing the UAV U i Whether to execute task T j , X ij =0 means UAV U i Do not perform task T j , X ij =1 means UAV U i Execute Task T j ; M represents the total number of drones, and N represents the total number of missions; The task completion time F time Modeled as: Among them, Δt is the UAV U i The time it takes to perform the task, For UAV i flight speed.

3. The method for allocating unmanned aerial vehicle cluster tasks based on an improved multi-objective ant colony algorithm according to claim 1 is characterized in that: In step 2, the coordination constraints between the drones in the drone cluster are expressed as: Where, X ij is the task decision variable, representing the UAV U i Whether to execute task T j , X ij =0 means UAV U i Do not perform task T j , X ij =1 means UAV U i Execute Task T j ; M represents the total number of drones, and N represents the total number of missions; This coordination constraint means that a UAV can only perform one task at a time, and a task can only be performed by one UAV; For any two adjacent tasks T that the drone is to perform l With task T j , the drone must complete the mission T l After that, task T can be executed j ,Therefore, the timing constraint of executing the task is expressed as: t j >t l +Δt Among them, t l ,t j They are respectively the UAV arrival task T l 、Task T j time, and task T l The execution order of tasks T j Before, Δt is the UAV U i The time it takes to perform the task; The flight range constraint of each UAV is expressed as: in, For UAV i The maximum flight range, d ij For UAV i In the execution of task T j The flight range is estimated by using the three-dimensional range cost calculation method based on the vertical section.

4. The method for allocating unmanned aerial vehicle cluster tasks based on an improved multi-objective ant colony algorithm according to claim 1 is characterized in that: In step 6, each drone competes for the right to execute the next task based on its member pheromones, which is expressed as: According to member pheromone gamma i (t), UAV U i The probability of obtaining the task execution right of the next task at time t is: Where M represents the total number of drones.

5. The method for allocating unmanned aerial vehicle cluster tasks based on an improved multi-objective ant colony algorithm according to claim 1 is characterized in that: In step 9, the updating of local pheromones is expressed as: in, is task T at time t+1 l With task T j The pheromone concentration between is the local pheromone volatility coefficient, and τ0 is the initial value of the pheromone.

6. The method for allocating unmanned aerial vehicle cluster tasks based on an improved multi-objective ant colony algorithm according to claim 1 is characterized in that: In step 11, the global pheromone of the two ant groups is updated, which is expressed as: in, The value of is non-negative; ρ is the global pheromone volatility coefficient; After each iteration, task T l With task T j The pheromone increment between is calculated as: In the formula, The pheromones left by all solutions in the archive set through the path (l, j); is the pheromone left by the solution in solution set Origin through path (l,j); Q is the pheromone constant; Origin is the solution set Origin; and They are the objective function values ​​of the solutions in the archive set and the objective function values ​​of the solutions in the solution set Origin respectively.

7. The method for allocating unmanned aerial vehicle cluster tasks based on an improved multi-objective ant colony algorithm according to claim 1 is characterized in that: In step 10, after the task allocation solutions corresponding to all ant clusters are obtained, q0 in the state transition rule is adaptively updated according to the following formula: Among them, q 0_max is the upper limit of q0, H is the population diversity set based on Shannon entropy; p ij is the probability of path (l, j) appearing in the task allocation solution; n a is the population size; E ij is the number of times the path (l, j) appears in the task allocation solution.

Images