A task allocation method for heterogeneous drone swarms based on coalition formation game

Abstract

The present invention discloses a method for allocating tasks for heterogeneous drone swarms based on an alliance formation game, comprising the following steps: Step 1: defining the collaborative task execution efficiency of drone swarms for two types of tasks, collaborative monitoring and collaborative evaluation; Step 2: modeling the drone swarm collaborative task execution problem as an alliance formation game model, where the game participants are the drones in the swarm; Step 3: solving the problem using an alliance formation algorithm based on partial cooperation to maximize the task execution efficiency of each drone and obtain a task selection result for each drone; Step 4: allocating drone swarm tasks based on the task selection result, thereby completing task allocation for heterogeneous drone swarms based on an alliance formation game. The present invention utilizes a game-based approach to enable each drone to have autonomous decision-making capabilities, with high timeliness and minimal information exchange. This method can effectively improve the autonomy of the drone swarm and maximize the task execution efficiency of the swarm.

Description

Technical Field

[0001] The present invention relates to a method for allocating tasks of a swarm of unmanned aerial vehicles (UAVs), and in particular to a method for allocating tasks of a swarm of heterogeneous UAVs based on alliance formation game. Background Art

[0002] Drones (UAVs) are widely used in various fields due to their high dynamism and dynamic deployment. With the rapid development of UAV technology, swarm control, and artificial intelligence, collaborative task execution by heterogeneous UAV swarms can effectively address the low efficiency and poor robustness of individual UAV tasks. For example, heterogeneous UAV swarms can perform multi-target monitoring and assessment tasks.

[0003] Currently, drone swarm task allocation can be broadly categorized into centralized and distributed approaches. Centralized approaches can effectively address the multi-constrained drone swarm task allocation problem, offering the advantage of providing a theoretically optimal solution. However, centralized approaches suffer from high computational complexity and are unsuitable for large-scale drone swarm task allocation. Furthermore, centralized approaches require global information, making them difficult to apply in highly dynamic and complex environments where information exchange between drones is unreliable. Distributed approaches primarily encompass some convex optimization algorithms and multi-agent reinforcement learning algorithms. Convex optimization algorithms are often limited by the structural characteristics of the problem, while multi-agent reinforcement learning algorithms face difficulties in guaranteeing convergence, training time, and storage constraints. Therefore, task allocation for heterogeneous drone swarms remains a highly challenging and hot topic. Summary of the Invention

[0004] Purpose of the invention: The technical problem to be solved by the present invention is to address the deficiencies of the existing technology and provide a task allocation method for a heterogeneous drone swarm based on alliance formation game.

[0005] In order to solve the above technical problems, the present invention discloses a method for allocating tasks of a heterogeneous drone swarm based on a coalition formation game, including the following steps:

[0006] Step 1: Define the collaborative task execution efficiency of the drone swarm according to the drone mission type;

[0007] The mission types of the UAV include collaborative monitoring and collaborative evaluation.

[0008] The aforementioned definitions of the UAV swarm collaborative task execution efficiency specifically include:

[0009] Step 1-1: When the UAV swarm performs the collaborative monitoring task type, calculate the task execution efficiency P of the UAV swarm R (S m ), the specific method is as follows:

[0010] When the drone swarm S mWhen executing collaborative task m, and the task is a collaborative monitoring task type, the task execution efficiency P of the drone group is R (S m ) is defined as:

[0011]

[0012] in, is the completion probability of UAV n for monitoring task m, n∈S m Indicates that drone n belongs to drone swarm S m , V m is the value coefficient of collaborative task m, R m is the threat coefficient of collaborative task m.

[0013] Step 1-2: When the UAV swarm performs the collaborative evaluation task type, calculate the task execution efficiency P of the UAV swarm. C (S m ), the specific method is as follows:

[0014] When the drone swarm S m When executing collaborative task m, and the task is a collaborative evaluation task type, the task execution efficiency P of the drone swarm is C (S m ) is defined as:

[0015]

[0016] in, is the completion probability of UAV n for evaluation task m.

[0017] Step 2: Model the UAV swarm collaborative task execution problem as a coalition formation game model, where the game participants are the UAVs in the swarm;

[0018] The problem of UAV group collaborative task execution is modeled as a coalition formation game model. as follows:

[0019]

[0020] in, Indicates the serial number of N drones in the drone group performing the mission, Represents the set of M tasks that the drone needs to perform, is the action set of N drones, where a n Action selection for drone n; is the utility function set of N drones, where u n is the utility function of UAV n, Alliance partition, that is, the alliance state formed by the drone swarm.

[0021] The drone utility function is as follows:

[0022] Using the average distribution criterion, define the utility function u of drone n n for:

[0023]

[0024] Among them, |S m | is the number of UAVs performing mission m.

[0025] Step 3: Use the partial cooperation-based coalition formation algorithm to maximize the task execution efficiency of each drone and obtain the task selection results of each drone. The specific steps are as follows:

[0026] Step 3-1, Initialization: Each drone randomly selects a task to form the initial alliance partition

[0027] Step 3-2: Randomly select a drone n, keep the task selection of other drones unchanged, and calculate the current task execution efficiency U of drone n n , the specific method is as follows:

[0028]

[0029] Among them, u j is the utility function of UAV j, For drone n to join the alliance, is the alliance of drone n before it joins. The symbol “\” indicates that the set element is removed from the set.

[0030] Step 3-3, drone n randomly selects a different task from task a n Mission According to step 3-2, calculate the task execution efficiency U′ of UAV n after changing the task selection. n ;

[0031] Step 3-4, if U n <U′ n , then drone n will select its mission from mission a n Change to a′ n ;if U n ≥U′ n , then UAV n maintains its mission selection a n constant;

[0032] Step 3-5: When convergence to a stable alliance partition, output the current task selection and task execution efficiency of each drone; otherwise, repeat step 3-2 until convergence to a stable alliance partition.

[0033] The convergence to the stable alliance partition is the sum of the mission execution efficiency of all drones Remains constant across iterations.

[0034] Step 4: Based on the above task selection results, assign the drone swarm tasks to complete the heterogeneous drone swarm task allocation based on the alliance formation game.

[0035] Beneficial effects:

[0036] By leveraging coalition game theory, this method empowers each drone with autonomous decision-making capabilities, making it more suitable for task allocation within large drone swarms. Furthermore, it eliminates the need for global information exchange, significantly reducing computational complexity. Compared to other distributed optimization methods, this game-based optimization approach offers high timeliness and minimal information exchange, effectively enhancing the autonomy of drone swarms and maximizing their mission execution efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more apparent.

[0038] Figure 1 It is a schematic diagram of the overall structure of the task allocation method of the present invention.

[0039] Figure 2 This is a flow chart of the alliance formation algorithm based on partial cooperation proposed in the present invention.

[0040] Figure 3 This is a graph showing the convergence simulation results of the partial cooperation-based alliance formation algorithm proposed in this invention. DETAILED DESCRIPTION

[0041] The present invention provides a task allocation method for a heterogeneous UAV swarm based on alliance formation game, so that the heterogeneous UAV swarm can perform heterogeneous task collaboration through local information interaction and exert cluster combat effectiveness.

[0042] Technical solution:

[0043] A task allocation method for a heterogeneous UAV swarm based on a coalition formation game includes the following steps:

[0044] Step 1: Define the collaborative task execution efficiency of drone swarms for collaborative monitoring and collaborative evaluation tasks respectively.

[0045] Step 2: Model the UAV swarm collaborative task execution problem as a coalition formation game model, where the game participants are the UAVs in the swarm.

[0046] Step 3: Use the partial cooperation-based coalition formation algorithm to solve the problem and maximize the mission execution efficiency of each drone.

[0047] In step 1, when the drone group S m When executing collaborative monitoring task m, its task execution efficiency P R (S m ) is defined as:

[0048]

[0049] in, is the completion probability of UAV n for task m, n∈S m Indicates that drone n belongs to swarm S m , V m is the value coefficient of task m, R m The threat coefficient of mission m. Similarly, when the drone group S m When executing collaborative evaluation task m, its task execution efficiency P C (S m ) is defined as:

[0050]

[0051] in, is the estimated task completion probability of UAV n for task m.

[0052] In step 2, the problem of cooperative task execution of drone swarms is modeled as a coalition formation game model, which is defined as:

[0053]

[0054] in, Indicates the serial number of N drones in the drone group performing the mission, Represents the set of M tasks that the drone needs to perform, is a set of actions, where a n Mission selection for UAV n; is the utility function, where u n is the utility function of UAV n, is the alliance partition, that is, the alliance state formed by the drone group. Using the average distribution criterion, the utility function of drone n is defined as:

[0055]

[0056] Among them, |S m | is the number of UAVs performing mission m.

[0057] In step 3, a coalition formation algorithm based on partial cooperation is proposed to maximize the mission execution efficiency of each drone. The specific algorithm is as follows:

[0058] Step 3-1, Initialization: Each drone randomly selects a task to form the initial alliance partition

[0059] Step 3-2: Randomly select a drone n, and keep the task selection of other drones unchanged. Calculate the current task execution efficiency U of drone n according to the following formula: n :

[0060]

[0061] Among them, u j is the utility function of UAV j, For drone n to join the alliance, is the alliance of drone n before it joins. The symbol “\” indicates that the set element is removed from the set.

[0062] Step 3-3, drone n randomly selects a different n Mission According to the above formula, the execution efficiency U′ of UAV n task after changing the task selection is calculated. n .

[0063] Step 3-4, if U n <U′ n , then UAV n will select its mission from a n Change to a′ n ;if U n ≥U′ n , then UAV n maintains its mission selection a n constant.

[0064] Step 3-5: When convergence to a stable alliance partition, output the current task selection and task execution efficiency of each drone; otherwise, repeat step 3-2 until convergence to a stable alliance partition.

[0065] Among them, convergence to a stable alliance partition means that no drone is willing to change its mission selection and the current alliance structure is fixed. The judgment standard is: the sum of the mission execution efficiency of all drones Remains constant across multiple iterations.

[0066] The principle of the present invention is:

[0067] This paper comprehensively considers mission value and mission threat coefficient to define mission execution efficiency for different tasks. Using game theory, this paper treats each drone in a swarm as a participant in the game. It uses a coalition formation game to solve the problem of collaborative mission execution in drone swarms. A partial cooperation-based coalition formation algorithm is proposed to solve this game model, thereby maximizing the mission execution efficiency of each drone.

[0068] Example:

[0069] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0070] The present invention provides a task allocation method for heterogeneous drone swarms based on alliance formation game, such as Figure 1 As shown, the following steps are included:

[0071] Step 1: Define the collaborative task execution efficiency of drone swarms for collaborative monitoring tasks and collaborative assessment tasks respectively.

[0072] When the drone swarm S m When executing collaborative monitoring task m, its task execution efficiency P R (S m ) is defined as:

[0073]

[0074] in, is the completion probability of UAV n for task m, n∈S m Indicates that drone n belongs to swarm S m , V m is the value coefficient of task m, R m The threat coefficient of mission m. Similarly, when the drone group S M When executing collaborative evaluation task m, its task execution efficiency P C (S m ) is defined as:

[0075]

[0076] in, is the estimated task completion probability of UAV n for task m.

[0077] Step 2: Model the UAV swarm collaborative task execution problem as a coalition formation game model, where the game participants are the UAVs in the swarm.

[0078] The game model is defined as:

[0079]

[0080] in, Indicates the serial number of N drones in the drone group performing the mission, Represents the set of M tasks that the drone needs to perform, is a set of actions, where a n Select the mission for UAV n; is the utility function, where u n is the utility function of UAV n, is the alliance partition, that is, the alliance state formed by the drone group. Using the average distribution criterion, the utility function of drone n is defined as:

[0081]

[0082] Among them, |S m | is the number of UAVs performing mission m.

[0083] Step 3: Use the partial cooperation-based coalition formation algorithm to solve the problem and maximize the mission execution efficiency of each drone.

[0084] like Figure 2 As shown in Figure 2, the steps of the alliance formation algorithm based on partial cooperation are as follows:

[0085] Step 3-1, Initialization: Each drone randomly selects a task to form the initial alliance partition

[0086] Step 3-2: Randomly select a drone n, and keep the task selection of other drones unchanged. Calculate the current task execution efficiency U of drone n according to the following formula: n :

[0087]

[0088] Among them, u j is the utility function of UAV j, For drone n to join the alliance, is the alliance of drone n before it joins. The symbol “\” indicates that the set element is removed from the set.

[0089] Step 3-3, drone n randomly selects a different n Mission According to the above formula, the execution efficiency U′ of UAV n task after changing the task selection is calculated. n .

[0090] Step 3-4, if U n <U′ n , then UAV n will select its mission from a n Change to a′ n ;if Un ≥U′ n , then UAV n maintains its mission selection a n constant.

[0091] Step 3-5: When convergence to a stable alliance partition, output the current task selection and task execution efficiency of each drone; otherwise, repeat step 3-2 until convergence to a stable alliance partition.

[0092] The task allocation method for heterogeneous drone swarms based on alliance formation game designed in this invention can effectively improve the task execution efficiency of drone swarms by applying it to specific examples. The specific applications are as follows:

[0093] Consider a scenario involving two monitoring tasks and two evaluation tasks, with a swarm of 10 drones of heterogeneous capabilities performing the different tasks. If a drone is solely a monitoring drone, its evaluation task completion probability is 0, and the monitoring task completion probability follows a uniform distribution over [0.4, 0.8]. If a drone is solely an evaluation drone, its monitoring task completion probability is also 0, and the evaluation task completion probability follows a uniform distribution over [0.4, 0.8]. If a drone is a dual-task drone, its monitoring task completion probability and the evaluation task completion probability follow a uniform distribution over [0.4, 0.8], respectively. Furthermore, the task value coefficient matrix is ​​[9.5, 8.4, 9.1, 7.6], and the task threat coefficient matrix is ​​[2.7, 2.2, 1.2, 2.8]. Figure 3 The results of the experiment are shown in Figure 2.

[0094] like Figure 3 As shown in the figure, it is a schematic diagram comparing the convergence of the drone swarm task execution efficiency under the application of the proposed alliance formation algorithm based on partial cooperation and the traditional alliance formation algorithm based on the Pareto criterion and selfish criterion. It can be seen that the convergence speed corresponding to the method designed by the present invention is the fastest, and the drone swarm task efficiency is significantly higher than the other two traditional methods.

[0095] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium is capable of storing a computer program that, when executed by the data processing unit, can execute the invention provided by the present invention, in particular, a method for allocating tasks to a heterogeneous drone swarm based on a coalition formation game, and some or all of the steps in each embodiment. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0096] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, in essence or in other words, the part that contributes to the prior art, can be embodied in the form of a computer program, i.e., a software product. The computer program software product can be stored in a storage medium and includes a number of instructions for enabling a device including a data processing unit (which can be a personal computer, server, single-chip microcomputer, MUU or network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0097] The present invention provides a concept and method for allocating tasks to a swarm of heterogeneous drones, specifically based on a coalition-forming game. There are numerous methods and approaches for implementing this technical solution. The foregoing is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art may make several improvements and modifications without departing from the principles of the present invention, and such improvements and modifications are also within the scope of protection of the present invention. Any components not specified in this embodiment may be implemented using existing technologies.

Claims

1. A task allocation method for heterogeneous drone swarms based on alliance formation game, characterized by: The steps include: Step 1: Define the collaborative task execution efficiency of the drone swarm according to the drone mission type; Step 2: Model the UAV swarm collaborative task execution problem as a coalition formation game model, where the game participants are the UAVs in the swarm; Step 3: Use the partial cooperation-based coalition formation algorithm to solve the problem, so as to maximize the task execution efficiency of each UAV and obtain the task selection result of each UAV; Step 4: Based on the above task selection results, assign the drone swarm tasks to complete the task allocation of the heterogeneous drone swarm based on the alliance formation game; The mission types of the UAVs described in step 1 include collaborative monitoring and collaborative assessment. The definition of the UAV swarm collaborative task execution efficiency described in step 1 includes: Step 1-1: When the UAV swarm performs the collaborative monitoring task type, calculate the task execution efficiency P of the UAV swarm R (S m ); Step 1-2: When the UAV swarm performs the collaborative evaluation task type, calculate the task execution efficiency P of the UAV swarm. C (S m ); The calculation of the task execution efficiency P of the drone swarm described in step 1-1 R (S m ), the specific method is as follows: When the drone swarm S m When executing collaborative task m, and the task is a collaborative monitoring task type, the task execution efficiency P of the drone group is R (S m ) is defined as: in, is the completion probability of UAV n for monitoring task m, n∈S m Indicates that drone n belongs to drone swarm S m , V m is the value coefficient of collaborative task m, R m is the threat coefficient of collaborative task m.

2. The method for allocating tasks of a heterogeneous drone swarm based on alliance formation game according to claim 1 is characterized in that: The calculation of the task execution efficiency P of the drone swarm described in step 1-2 C (S m ), the specific method is as follows: When the drone swarm S m When executing collaborative task m, and the task is a collaborative evaluation task type, the task execution efficiency P of the drone swarm is C (S m ) is defined as: in, is the completion probability of UAV n for evaluation task m.

3. The method for allocating tasks of a heterogeneous drone swarm based on alliance formation game according to claim 2 is characterized in that: The problem of cooperative task execution of drone groups described in step 2 is modeled as a coalition formation game model. as follows: in, Indicates the serial number of N drones in the drone group performing the mission, Represents the set of M tasks that the drone needs to perform, is the action set of N drones, where a n Action selection for drone n; is the utility function set of N drones, where u n is the utility function of UAV n, Alliance partition, that is, the alliance state formed by the drone swarm.

4. The method for allocating tasks of a heterogeneous drone swarm based on alliance formation game according to claim 3 is characterized in that: The drone utility function described in step 2 is as follows: Using the average distribution criterion, define the utility function u of drone n n for: Among them, |S m | is the number of UAVs performing mission m.

5. The method for allocating tasks of a heterogeneous drone swarm based on alliance formation game according to claim 4 is characterized in that: The specific steps of the partial cooperation-based alliance formation algorithm described in step 3 are as follows: Step 3-1, Initialization: Each drone randomly selects a task to form the initial alliance partition Step 3-2: Randomly select a drone n, keep the task selection of other drones unchanged, and calculate the current task execution efficiency U of drone n n ; Step 3-3, drone n randomly selects a different task from task a n Mission Among them, \ represents the set element a n From this collection Eliminate the task and calculate the task execution efficiency U′ of UAV n after changing the task selection according to step 3-2 n ; Step 3-4, if U n <U′ n , then drone n will select its mission from mission a n Change to a′ n ;if U n ≥U′ n , then UAV n maintains its mission selection a n constant; Step 3-5: When convergence to a stable alliance partition, output the current task selection and task execution efficiency of each drone; otherwise, repeat step 3-2 until convergence to a stable alliance partition.

6. The method for allocating tasks of a heterogeneous drone swarm based on alliance formation game according to claim 5 is characterized in that: The calculation of the current task execution efficiency U of drone n described in step 3-2 n , the specific method is as follows: Among them, u j is the utility function of UAV j, For drone n to join the alliance, For drone n before joining the alliance.

7. The method for allocating tasks of a heterogeneous drone swarm based on alliance formation game according to claim 6 is characterized in that: Converge to a stable alliance partition as described in steps 3-5, that is, the sum of the mission execution efficiency of all drones Remains constant across iterations.

Images