Cluster distributed dynamic target allocation method based on preference alliance game

Abstract

The invention discloses a cluster distributed dynamic target allocation method based on a preference alliance game, and the method comprises the steps: defining a decision variable matrix to represent a matching relation between a target and an aircraft, defining a cluster aircraft target conflict constraint and a target matching constraint condition, and building a cluster communication topology; constructing a strike efficiency function based on the probability that the aircraft strikes the target successfully, designing an aircraft revenue function, and establishing an aircraft optimization performance index; designing a preference-based target selection rule based on the revenue function and the optimization performance index of the aircraft, and establishing a preference alliance game model; based on a preference alliance game model and a maximum information gain algorithm, a distributed dynamic target allocation algorithm is designed, so that a cluster aircraft reaches a Nash stable partition through local information interaction under a communication topology. The method can solve the dynamic target distribution problem of a large-scale cluster, and has the advantages of small communication load, high adaptability and the like.

Description

Technical Field

[0001] The present invention belongs to the field of distributed task allocation, and in particular relates to a cluster distributed dynamic target allocation method based on preference alliance game. Background Art

[0002] Swarm aircraft is a system composed of a large number of aircraft of different types and simple structures. Through the functional complementarity and collaboration between aircraft, swarm aircraft can adapt to complex and diverse mission scenarios such as emergency rescue and reconnaissance operations, improve the efficiency and success rate of mission execution, and have broad application prospects.

[0003] Target allocation is one of the core modules of the decision-making layer of swarm aircraft. As an action guide for executing tasks, it aims to achieve matching between aircraft and targets by analyzing aircraft, targets, and environmental information, providing an important basis for subsequent trajectory planning and individual aircraft control. However, due to the large scale of swarm aircraft and the frequent failure of individual aircraft, it is difficult to rely on operators to directly allocate and adjust the targets of each aircraft. Therefore, the study of distributed dynamic target allocation methods for swarms is of great significance to expanding the scale of swarms and improving the reliability of swarms.

[0004] The cluster distributed dynamic target allocation method needs to meet the following requirements: (1) Information is locally known, that is, individual aircraft can only obtain information about some related aircraft and cannot grasp the information of the entire cluster. (2) Individual independent decision-making, that is, individual aircraft needs to select the optimal target based on the local information obtained and update the target that matches itself. (3) The convergence of the algorithm, that is, through the target selection and target update of individual aircraft, the target allocation result of the cluster can tend to remain unchanged. (4) The fault tolerance of the algorithm, that is, when some individual aircraft fail, can the cluster autonomously optimize and adjust the target allocation result of the cluster through the independent decision-making of individual aircraft. The existing target allocation methods are difficult to meet the above requirements. Therefore, in response to the above requirements, the present invention designs a cluster distributed dynamic target allocation method based on the preference coalition game theory. Summary of the invention

[0005] Purpose of the invention: The present invention provides a cluster distributed dynamic target allocation method based on preference coalition game, which can solve the real-time distributed target allocation problem of large-scale clusters, reduce cluster communication load, and improve cluster reliability.

[0006] Technical solution: The cluster distributed dynamic target allocation method based on preference alliance game described in the present invention includes the following steps:

[0007] (1) Define the decision variable matrix to characterize the matching relationship between the target and the aircraft, clarify the target conflict constraints and target matching constraints of the cluster aircraft, and establish the cluster communication topology;

[0008] (2) Construct a strike effectiveness function based on the probability of the aircraft successfully striking the target, and then design the aircraft benefit function and establish the aircraft optimization performance index;

[0009] (3) Design preference-based target selection rules based on the aircraft's revenue function and optimization performance indicators, and establish a preference alliance game model;

[0010] (4) Based on the preference coalition game model and the maximum information gain algorithm, a distributed dynamic target allocation algorithm is designed so that cluster aircraft can reach the Nash stable partition through local information interaction under the communication topology.

[0011] Furthermore, the decision variable matrix defined in step (1) to characterize the matching relationship between the target and the aircraft is:

[0012] Given p (targets and m aircraft, p>m); define the decision variable matrix A = [α ij ] m×(p+1) , where target 0 corresponds to standby, α ij ∈{0,1} represents the aircraft i∈M and the target The matching relationship, α ij =1 indicates a successful match; otherwise, the match fails.

[0013] Furthermore, the target conflict constraint and target matching constraint of the cluster aircraft in step (1) are:

[0014] Considering that each aircraft can only attack one target, establish the target conflict constraint:

[0015]

[0016] Considering the limited attack range and maneuverability of the aircraft, according to the coordinate information of the aircraft i∈M Coordinate information of target j∈P The center coordinates of the no-fly zone l∈Q Coverage radius R l , establish target matching constraints:

[0017]

[0018]

[0019] Q ij ={l∈Q|d ijl <R l}

[0020] Among them, D ij The comprehensive aircraft's long-range strike capability and maneuverability, λ is a positive constant, ρ i It is a parameter that characterizes the long-range strike capability of an aircraft. It is a parameter that characterizes the maneuverability of the aircraft; d ij is the straight-line distance from aircraft i to target j; is the distance d at which the line connecting aircraft i and target j overlaps the no-fly zone ijl is the vertical distance from the center coordinate of the no-fly zone l to the line connecting the aircraft i and the target j, Q ij It is the set of no-fly zones that aircraft i needs to avoid when flying to target j.

[0021] Furthermore, the cluster communication topology in step (1) is:

[0022] According to the target matching constraint, the real target set that the aircraft i∈M can select is defined as An undirected graph G with aircraft sets as nodes is established to represent the communication topology of the aircraft cluster; the adjacency matrix of the undirected graph G is B = [β ik ] m×m , where if P i ∩P k ≠φ and i≠k, then β ik =β ki =1; otherwise β ik =β ki =0; definition is the neighbor set of aircraft i.

[0023] Furthermore, the implementation process of step (2) is as follows:

[0024] The closer the distance between aircraft i and target j is, the shorter the overlap area between the line connecting the two and the no-fly zone is, and the stronger the strike efficiency of the aircraft is. i The strike effectiveness function of aircraft i against target j∈P i The probability function of a successful attack is as follows:

[0025]

[0026] Defines the currently selected target The set of aircraft is Λ j ,Target The benefit of being successfully attacked is given by the set Λ j Internal aircraft share, aircraft i selects the target The profit function is:

[0027]

[0028] The local optimization objective of aircraft i is established as follows:

[0029]

[0030] in, is the optimization performance index of aircraft i, s i The target selected for vehicle i, The set of targets selected for the neighbors of aircraft i.

[0031] Furthermore, the probability of the aircraft successfully striking the target in step (2) also includes the probability of multiple aircraft successfully striking target j:

[0032]

[0033] Among them, Λ j Select the target for the current The set of aircraft, u ij is the probability of a single aircraft i successfully hitting target j.

[0034] Furthermore, the probability of successfully hitting target 0 is set to 1.

[0035] Furthermore, the implementation process of step (3) is as follows:

[0036] Definition Symbols> i Characterize the preference relationship between aircraft i∈M and the alliance. If for target j, μ∈P i ∪{0}, satisfies:

[0037]

[0038] Then Λ j > i Λ μ , indicating that aircraft i has a higher probability of joining the alliance Λ μ , prefer to join the alliance j ;

[0039] Based on this preference relationship, the target selection rule of the aircraft is established as follows: Given a coalition partition Ω = {Λ0, Λ1, ... Λ p}, for any aircraft i∈M, if and only if there is a target and Make The aircraft will then send the currently matched target i Adjust to target From the Alliance Adjust to Alliance

[0040] Based on the aircraft set M and target set As well as the above preference-based target selection rules, a preference alliance game model is established.

[0041] Furthermore, the distributed dynamic target allocation algorithm in step (4) includes target selection and information sending and information receiving and target updating, wherein:

[0042] Target selection and information sending are:

[0043] 1) Introducing variables Characterize the target selected by aircraft i, and set its initial value to s i , introduce the collection To avoid aircraft i from repeatedly selecting the same target, its initial value is set to P i ∪{0}\{s i};

[0044] 2) Judgment set Is it an empty set? If it is an empty set, jump to 3); if it is not an empty set, then in the set Select any target If satisfied This means that aircraft i is more likely to join the alliance Prefer to join an alliance Therefore, Assign to and will From the collection Remove from the list and jump to 2);

[0045] 3) Calculate the number of aircraft i by the alliance Adjust to Alliance Post-optimization performance index J i The increase of ΔJ i , ΔJ i Send to all neighbors of aircraft i; when When the increase is ΔJ i is 0;

[0046] Information received and target updated as:

[0047] Based on the received information about the changes in the performance indicators of the aircraft in the neighbor set, determine whether the If satisfied, it means that the current aircraft i changes the target selection to After that, the increase of its performance index is greater than that of its neighbors. Then, based on the idea of ​​the maximum information gain algorithm, this iteration ends and the aircraft i changes its target selection to If it is not satisfied, it means that there is a neighbor whose performance index increase is greater than or equal to that of aircraft i. Then, this iteration ends and aircraft i still chooses target s.i .

[0048] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are: the present invention can not only effectively solve the problem of dynamic target allocation of large-scale cluster aircraft, but also has significant advantages such as small communication load and strong adaptability; the present invention significantly improves the collaborative combat capability and task execution efficiency of cluster aircraft through intelligent game mechanism and efficient algorithm design, and has high application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of the present invention;

[0050] Figure 2 A schematic diagram of an aircraft target matching constraint according to an embodiment of the present invention;

[0051] Figure 3 A flow chart of a cluster distributed dynamic target allocation algorithm according to an embodiment of the present invention;

[0052] Figure 4 To simulate the target allocation results of the cluster in a fault-free situation on the experimental platform;

[0053] Figure 5 This is a curve diagram of the change of cluster benefits when simulating a fault-free situation on the experimental platform;

[0054] Figure 6 To simulate the target allocation results of the cluster under the condition of aircraft failure on the experimental platform;

[0055] Figure 7 This is a curve diagram of the change in cluster benefits when simulating aircraft failure on the experimental platform. DETAILED DESCRIPTION

[0056] The present invention is further described in detail below with reference to the accompanying drawings.

[0057] like Figure 1 As shown, the present invention provides a cluster distributed dynamic target allocation method based on preference alliance game, which specifically includes the following steps:

[0058] Step 1: Define the decision variable matrix to characterize the matching relationship between the target and the aircraft, clarify the target conflict constraints and target matching constraints of the cluster aircraft, and establish the cluster communication topology.

[0059] The target allocation scenario considered in this implementation is as follows: Figure 2 As shown, including aircraft, targets, and no-fly zones; among them:

[0060] Aircraft: m aircraft, define the aircraft set as The coordinate information of aircraft i∈M is recorded as

[0061] Target: p (p>m) targets, the target set is defined as Among them, the set P = {1, 2, ..., p} corresponds to the real target, and the target 0 corresponds to standby, that is, it does not match any target. The coordinate information of the target j∈P is recorded as The value is V j , that is, the benefit obtained after the target is successfully hit. In particular, the value of standby is 0.

[0062] No-fly zones: There are q no-fly zones, and the set of no-fly zones is defined as The center coordinates of the no-fly zone l∈Q are The coverage radius is denoted as R l Neither the aircraft nor the target is within the no-fly zone.

[0063] Consider assigning p (p>m) targets to m aircraft. In order to characterize the matching relationship between targets and aircraft, the decision variable matrix is ​​defined as Among them, α ij ∈{0,1} represents the aircraft i∈M and the target The matching relationship, where α ij =1 indicates that the aircraft i and target j are matched successfully; otherwise, the aircraft i and target j are matched unsuccessfully.

[0064] Considering that each aircraft can only attack one target, the following target conflict constraints are established:

[0065]

[0066] Due to the limited strike range and maneuverability of the aircraft, when selecting a target, it is necessary not only to consider whether the straight-line distance from the aircraft to the target meets the strike range, but also to consider whether the aircraft can avoid the no-fly zone. In order to reflect this limitation, for aircraft i∈M and target j∈P, the following target matching constraints are established:

[0067]

[0068] Among them, D ij The comprehensive aircraft long-range strike capability and maneuverability is in the form of a weighted sum; λ is a positive constant, ρ i is a parameter that characterizes the long-range strike capability of an aircraft, d ij is the straight-line distance from aircraft i to target j, is a parameter that characterizes the maneuverability of the aircraft. is the distance at which the line connecting aircraft i and target j overlaps with the no-fly zone, such as Figure 2 As shown; d ijlis the vertical distance from the center coordinate of the no-fly zone l to the line connecting the aircraft i and the target j, Q ij It is the set of no-fly zones that aircraft i needs to avoid when flying to target j.

[0069] According to the target matching constraint, the real target set that the aircraft i∈M can select is defined as In order to save communication resources and reduce communication interference, for any two aircraft i, k∈M, if and only if they can select the same real target, that is, P i ∩P k ≠φ, they can communicate with each other. Therefore, an undirected graph G with aircraft sets as nodes is established to represent the communication topology of the aircraft cluster, and the adjacency matrix of the undirected graph G is defined as B = [β ik ] m×m , where if P i ∩P k ≠φ and i≠k, then β ik =β ki =1, otherwise β ik =β ki = 0. Definition is the neighbor set of aircraft i.

[0070] Step 2: Construct a strike effectiveness function based on the probability of the aircraft successfully striking the target, then design the aircraft benefit function and establish the aircraft optimization performance index.

[0071] Aircraft i to target j∈P i The striking efficiency of the two is related to the distance d ij The overlap distance between the line connecting the two and the no-fly zone Related, ij The smaller it is, the greater the striking effectiveness; The smaller it is, the greater the strike effectiveness is. Based on this relationship, the relationship between aircraft i and target j∈P is established. i The strike effectiveness function of aircraft i against target j∈P i The probability function of a successful attack is as follows:

[0072]

[0073] Since the number of targets p is greater than the number of aircraft m, there are often situations where multiple aircraft cooperate to attack the same target. Define the current selected target The set of aircraft is Λ j , according to the target j∈P of aircraft i i The attack effectiveness function of the target j∈P i The probability of being successfully attacked is:

[0074]

[0075] Specifically, the probability of target 0 being successfully hit is set to 1.

[0076] If the aircraft set Λ j Successfully hit target j, then the target value V j is obtained by the collaboration of multiple aircraft and needs to be apportioned to Λ j Therefore, if aircraft i chooses target j, its profit function can be expressed as:

[0077]

[0078] Among them, |Λ j | represents alliance Λ j The number of aircraft inside.

[0079] It can be seen from the cluster communication topology that for any aircraft i∈M, it can only obtain information about neighboring aircraft. Therefore, the local optimization goal of aircraft i is established as follows:

[0080]

[0081] in, is the optimization performance index of aircraft i, s i The target selected for vehicle i, The target set selected for the neighbors of aircraft i. It can be seen that the local optimization goal of aircraft i is to maximize the total benefit of itself and the aircraft in the neighbor set.

[0082] Step 3: Design preference-based target selection rules based on the aircraft's revenue function and optimization performance indicators, and establish a preference coalition game model.

[0083] Select the target The set of aircraft Λ j Defined as Alliance Λ j , define the alliance partition of the cluster as Ω={Λ0,Λ1,…Λ p}. Define symbol > i Characterize the preference relationship between aircraft i∈M and alliances. For any two alliances Λ j , Λ μ , if j > i Λ μ ,j,μ∈P i ∪{0}, it means that aircraft i has a higher probability of joining the alliance Λ μ , prefer to join the alliance j Based on the revenue function and optimization goal of aircraft i, we can further give the preference relationship between aircraft and alliance: if for target j, μ∈Pi ∪{0}, satisfies:

[0084]

[0085] Then Λ j > i Λ μ Based on this preference relationship, the target selection rule of the aircraft can be established as follows: Given a coalition partition Ω = {Λ0, Λ1, ... Λ p}, for any aircraft i∈M, if and only if there is a target and Make The aircraft will then send the currently matched target i Adjust to target That is, from the alliance adjustment To Alliance The alliance partition is composed of Ω = {Λ0, Λ1, ... Λ p Updated to

[0086] According to the aircraft set M and target set As well as the above preference-based target selection rules, a preference coalition game model can be established. Accordingly, the target allocation problem can be transformed into a coalition partition problem, that is, finding a target set that matches the cluster aircraft. and its corresponding alliance division So that for any aircraft i∈M and target All meet Since each aircraft prefers the target it currently chooses in this partition, none of them will increase their benefits by changing their chosen targets. This kind of alliance division is called a Nash stable partition. Under the Nash stable partition, it is easy to get It is not difficult to see that this is consistent with the local optimization goal of aircraft i.

[0087] Step 4: Based on the preference coalition game model and the maximum information gain algorithm, a distributed dynamic target allocation algorithm is designed so that cluster aircraft can reach the Nash stable partition through local information interaction under the communication topology.

[0088] In order to make the cluster aircraft reach the Nash stable partition through local information interaction under the communication topology, such as Figure 3 As shown, the distributed dynamic target allocation algorithm corresponds to the decision-making idea of ​​aircraft i in one iteration and consists of two sub-steps.

[0089] Target selection and message sending:

[0090] 1) Introducing variables Characterize the target selected by aircraft i, and set its initial value to s i , introduce the collection To avoid aircraft i from repeatedly selecting the same target, its initial value is set to P i ∪{0}\{s i}.

[0091] 2) Judgment set Is it an empty set? If it is an empty set, jump to 3); if it is not an empty set, then in the set Select any target If satisfied This means that aircraft i is more likely to join the alliance Prefer to join an alliance Therefore, Assign to and will From the collection Remove from the list and skip to 2).

[0092] 3) Calculate the number of aircraft i by the alliance Adjust to Alliance * Post-optimization performance index J i The increase of ΔJ i (when When ΔJ i Send to all neighbors of aircraft i.

[0093] Information reception and target update. Based on the information received about the changes in the performance indicators of the aircraft in the neighbor set, determine whether the If satisfied, it means that the current aircraft i changes the target selection to After that, the increase of its performance index is greater than that of its neighbors. Then, based on the idea of ​​the maximum information gain algorithm, this iteration ends and the aircraft i changes its target selection to If it is not satisfied, it means that there is a neighbor whose performance index increase is greater than or equal to that of aircraft i. Then, this iteration ends and aircraft i still chooses target s. i .

[0094] In order to verify the effectiveness of the present invention, the present invention performs a no-fault situation simulation and a fault situation simulation.

[0095] The simulation of the fault-free situation is based on the Matlab simulation platform. Considering the task allocation scenario consisting of 10 aircraft and 4 targets, the coordinate information of each aircraft, the coordinates and importance information of each target, the center coordinates of each no-fly zone, and the coverage radius information are shown in Tables 1, 2, and 3.

[0096] Table 1 Coordinate information of each aircraft simulated by the experimental platform

[0097] Aircraft number Horizontal axis Vertical axis Aircraft number Horizontal axis Vertical axis 1 15 12 6 9 56 2 5 20 7 8 63 3 2 30 8 13 70 4 6 45 9 14 82 5 8 50 10 16 86

[0098] Table 2 Coordinates and importance information of each target simulated by the experimental platform

[0099]

[0100]

[0101] Table 3 Experimental platform simulates the center coordinates and coverage radius of each no-fly zone

[0102] No-fly zone number Center horizontal coordinate Center ordinate Coverage Radius 1 40 33 6

[0103] For aircraft i∈M, let ρ i =ρ=0.016, λ=1. At the initial moment, all aircraft are set to standby, and the task allocation results are as follows: Figure 4 As shown. Figure 4 It can be seen that the task allocation result of the drone is:

[0104] Target 1: {1,2,3}; Target 2: {4,5}; Target 3: {7,8,9}; Target 4: {6,10}.

[0105] Figure 5 A curve showing the variation of the total revenue of the cluster with the number of iterations is given. It can be seen that under the distributed task algorithm designed by the present invention, the total revenue of the cluster continues to increase and finally converges at the 11th iteration.

[0106] In the fault situation simulation, the distributed task allocation algorithm proposed in the present invention can adapt to the failure. Figure 3 It can be seen that if for any aircraft i∈M, after removing the failed aircraft, the current selected goal is not to make the performance index If the optimal goal of maximizing is reached, then in this iteration, at least one of aircraft i and its neighbor aircraft will change its goal. After that, the cluster will continue to iterate until the goal selected by each aircraft is the optimal goal that maximizes its own performance index.

[0107] In order to compare with the no-failure situation, the coordinate information of each aircraft, the coordinates and importance of each target, the center coordinates of each no-fly zone, and the coverage radius information are the same as Table 1, Table 2, and Table 3. The cluster is adjusted based on the task allocation results in the no-failure situation. Considering that the set of aircraft with complete failure is {6,10}, the task allocation results of the cluster after the failure are as follows Figure 6 As shown. Figure 6 It can be seen that the task allocation result of the drone is:

[0108] Target 1: {1,2,3}; Target 2: {4,5}; Target 3: {7,8}; Target 4: {9}

[0109] Figure 7 The curve of the total revenue of the cluster changing with the number of iterations is given. It can be seen that the total revenue of the cluster decreases in the 12th iteration due to the occurrence of complete failure. After that, under the designed distributed task algorithm, the total revenue of the cluster continues to increase and finally converges in the 13th iteration.

[0110] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A cluster distributed dynamic target allocation method based on preference coalition game, characterized in that: The following steps are involved: (1) Define the decision variable matrix to characterize the matching relationship between the target and the aircraft, clarify the target conflict constraints and target matching constraints of the cluster aircraft, and establish the cluster communication topology; (2) Construct a strike effectiveness function based on the probability of the aircraft successfully striking the target, and then design the aircraft benefit function and establish the aircraft optimization performance index; (3) Design preference-based target selection rules based on the aircraft's revenue function and optimization performance indicators, and establish a preference alliance game model; (4) Based on the preference coalition game model and the maximum information gain algorithm, a distributed dynamic target allocation algorithm is designed so that cluster aircraft can reach the Nash stable partition through local information interaction under the communication topology.

2. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The decision variable matrix described in step (1) is defined to characterize the matching relationship between the target and the aircraft as follows: Given p targets and m aircraft, p>m); Define the decision variable matrix A = [α ij ] m×(p+1) , where target 0 corresponds to standby, α ij ∈{0,1} represents the aircraft i∈M and the target The matching relationship, α ij =1 indicates a successful match; otherwise, the match fails.

3. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The target conflict constraint and target matching constraint conditions of the cluster aircraft in step (1) are: Considering that each aircraft can only attack one target, establish the target conflict constraint: Considering the limited attack range and maneuverability of the aircraft, according to the coordinate information of the aircraft i∈M Coordinate information of target j∈P The center coordinates of the no-fly zone l∈Q Covering radius R l , establish target matching constraints: Among them, D ij The comprehensive aircraft's long-range strike capability and maneuverability, λ is a positive constant, ρ i It is a parameter that characterizes the long-range strike capability of an aircraft. It is a parameter that characterizes the maneuverability of the aircraft; d ij is the straight-line distance from aircraft i to target j; is the distance d at which the line connecting aircraft i and target j overlaps the no-fly zone ijl is the vertical distance from the center coordinate of the no-fly zone l to the line connecting the aircraft i and the target j, Q ij It is the set of no-fly zones that aircraft i needs to avoid when flying to target j.

4. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The cluster communication topology in step (1) is: According to the target matching constraint, the real target set that the aircraft i∈M can select is defined as An undirected graph G with aircraft sets as nodes is established to represent the communication topology of the aircraft cluster; the adjacency matrix of the undirected graph G is B = [β ik ] m×m , where if P i ∩P k ≠φ and i≠k, then β ik =β ki =1; otherwise β ik =β ki =0; definition is the neighbor set of aircraft i.

5. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The implementation process of step (2) is as follows: The closer the distance between aircraft i and target j is, the shorter the overlap area between the line connecting the two and the no-fly zone is, and the stronger the strike efficiency of the aircraft is. i The strike effectiveness function of aircraft i against target j∈P i The probability function of a successful attack is as follows: Define the set of aircraft that currently selects the target j∈P as Λ j ,Target The benefit of being hit successfully is given by the set Λ j Internal aircraft share, aircraft i selects the target The profit function is: The local optimization objective of aircraft i is established as follows: in, is the optimization performance index of aircraft i, s i The target selected for vehicle i, The set of targets selected for the neighbors of aircraft i.

6. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The probability of the aircraft successfully striking the target in step (2) also includes the probability of multiple aircraft successfully striking target j: Among them, Λ j Select the target for the current The set of aircraft, u ij is the probability of a single aircraft i successfully hitting target j.

7. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The probability of a successful hit on target 0 is set to 1.

8. According to claim 1, a cluster distributed dynamic target allocation method based on preference alliance game is characterized in that: The implementation process of step (3) is as follows: Definition Symbols> i Characterize the preference relationship between aircraft i∈M and the alliance. If for target j, μ∈P i ∪{0}, satisfies: Then Λ j > i Λ μ , indicating that aircraft i has a higher probability of joining the alliance Λ μ , prefer to join the alliance j ; Based on this preference relationship, the target selection rule of the aircraft is established as follows: Given a coalition partition Ω = {Λ0, Λ1, ... Λ p }, for any aircraft i∈M, if and only if there is a target and Make The aircraft will then send the currently matched target i Adjust to target From the Alliance Adjust to Alliance Based on the aircraft set M and target set As well as the above preference-based target selection rules, a preference alliance game model is established.

9. The cluster distributed dynamic target allocation method based on preference alliance game according to claim 1 is characterized in that: The distributed dynamic target allocation algorithm in step (4) includes target selection and information sending and information receiving and target updating, wherein: Target selection and information sending are: 1) Introducing variables Characterize the target selected by aircraft i, and set its initial value to s i , introduce the collection To avoid aircraft i from repeatedly selecting the same target, its initial value is set to P i ∪{0}\{s i }; 2) Judgment set Is it an empty set? If it is an empty set, jump to 3); if it is not an empty set, then in the set Select any target If satisfied This means that aircraft i is more likely to join the alliance Prefer to join an alliance Therefore, Assign to and will From the collection Remove from the list and jump to 2); 3) Calculate the number of aircraft i by the alliance Adjust to Alliance Post-optimization performance index J i The increase of ΔJ i , ΔJ i Send to all neighbors of aircraft i; Information received and target updated as: Based on the received information about the changes in the performance indicators of the aircraft in the neighbor set, determine whether the If satisfied, it means that the current aircraft i changes the target selection to After that, the increase of its performance index is greater than that of its neighbors. Then, based on the idea of ​​the maximum information gain algorithm, this iteration ends and the aircraft i changes its target selection to If it is not satisfied, it means that there is a neighbor whose performance index increase is greater than or equal to that of aircraft i. Then, this iteration ends and aircraft i still chooses target s. i .

10. The cluster distributed dynamic target allocation method based on preference alliance game according to claim 1 is characterized in that: Said When the increase is ΔJ i is 0.

Images