Heterogeneous aircraft cooperative task planning method based on improved wolf pack

Abstract

The invention relates to a heterogeneous aircraft cooperative task planning method, in particular to a heterogeneous aircraft cooperative task planning method based on an improved wolf pack. The invention provides a heterogeneous aircraft cooperative task planning method based on an improved wolf pack in order to solve the problem of low multi-task allocation accuracy of existing heterogeneous aircrafts. The heterogeneous aircraft cooperative task planning method based on the improved wolf pack specifically comprises the steps of 1, constructing a heterogeneous aircraft task optimal allocation model; and 2, optimizing the heterogeneous aircraft task optimal allocation model by adopting an improved wolf pack optimization algorithm to obtain an optimal task allocation scheme. The method is applied to the field of heterogeneous aircraft cooperative task planning.

Description

Technical Field

[0001] The present invention relates to a collaborative mission planning method for heterogeneous aircraft. Background Art

[0002] Existing technologies have not fully utilized the ability of aircraft to perform heterogeneous and diversified tasks, lack in-depth research on related heterogeneous mission planning issues, and insufficiently consider the various constraints of similar problems.

[0003] The wolf pack algorithm is simple and easy to execute, and is suitable for solving task allocation problems. However, it is highly dependent on the initial population distribution and uses a fixed step size in its wandering strategy. The optimization efficiency is unstable and it is easy to fall into local optimality.

[0004] In summary, the accuracy of multi-task allocation for existing heterogeneous aircraft is low. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem of low accuracy of multi-task allocation of existing heterogeneous aircraft and to propose a heterogeneous aircraft collaborative task planning method based on improved wolf pack.

[0006] The specific process of the collaborative mission planning method for heterogeneous aircraft based on improved wolf pack is as follows:

[0007] Step 1: Construct an optimal allocation model for heterogeneous aircraft tasks;

[0008] Step 2: Use the improved wolf pack optimization algorithm to optimize the optimal task allocation model of heterogeneous aircraft and obtain the optimal task allocation plan.

[0009] The beneficial effects of the present invention are:

[0010] This paper considers a large-capacity, long-range, high-speed aircraft that can carry multiple payloads or configure internal space for personnel rescue and material delivery missions, respectively. It proposes a method for aircraft collaborative task planning based on an improved wolf pack algorithm and establishes an optimal allocation model for heterogeneous aircraft performing rescue or material delivery missions, including a capability model, a reward model, and a constraint model. To address the problem that traditional wolf pack algorithms are prone to falling into local optimality, a chaotic mapping and reverse learning approach are proposed to achieve uniform population initialization. Furthermore, a Levy flight strategy is introduced, and a variable-step-size search strategy based on the Mantegna algorithm is proposed to escape local optimal solutions, fully utilizing the aircraft's ability to perform heterogeneous and diversified tasks and improving the accuracy of multi-task allocation for heterogeneous aircraft.

[0011] This invention relates to a method for collaborative mission planning for heterogeneous aircraft based on an improved wolf pack. Considering multiple aircraft with different mission execution capabilities, an optimal allocation model for heterogeneous collaborative flight missions is established. Furthermore, a population initialization method based on chaotic mapping and reverse learning, as well as a variable-step-size walk strategy based on Lévy flight, are proposed to form an improved wolf pack optimization algorithm. This algorithm fully utilizes the aircraft's ability to execute heterogeneous and diversified missions, and improves the accuracy of multi-tasking allocation for heterogeneous aircraft.

[0012] 1. Key technical points of the present invention

[0013] (1) Optimal multi-task allocation model for heterogeneous aircraft

[0014] Considering an aircraft with large payload, long range and high speed, which can carry multiple payloads or configure internal space for personnel rescue missions and material delivery missions respectively, an optimal allocation model for heterogeneous aircraft to perform rescue or material delivery missions is established, including capability model, reward model and constraint model.

[0015] (2) Improved wolf pack algorithm optimization algorithm

[0016] A population initialization method based on chaotic mapping and reverse learning and a variable step-size walking strategy based on Lévy flight are proposed to form an improved wolf pack optimization algorithm, which can give full play to the ability of the aircraft to perform heterogeneous and diversified tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is the flow chart of the improved wolf pack algorithm optimization algorithm of the present invention. DETAILED DESCRIPTION

[0018] Specific implementation method 1: The specific process of the heterogeneous aircraft collaborative mission planning method based on the improved wolf pack in this implementation method is as follows:

[0019] Step 1: Construct an optimal multi-task allocation model for heterogeneous aircraft;

[0020] Step 2: Use the improved wolf pack optimization algorithm to optimize the multi-task optimal allocation model of heterogeneous aircraft and obtain the optimal task allocation plan.

[0021] Specific embodiment 2: This embodiment differs from specific embodiment 1 in that: in step 1, a heterogeneous aircraft multi-task optimal allocation model is constructed; the specific process is as follows:

[0022] (1) Suppose N aircraft jointly perform M tasks;

[0023] The tasks are divided into rescue tasks and delivery tasks, and the types are represented by K i Indicates that K i =1 means rescue mission, K i=2 indicates the delivery task, i=1,2,…,M;

[0024] (2) Make the following assumptions about the aircraft task allocation problem:

[0025] 1) Terrain obstacles and sudden threats are not considered;

[0026] 2) Spatial distance is expressed as two-dimensional Euclidean distance;

[0027] 3) The destination corresponding to each task remains stationary;

[0028] 4) The aircraft flies at a constant speed;

[0029] 5) The aircraft selects the mission closest to its current location as the next mission;

[0030] 6) The aircraft has no communication distance or fuel restrictions, and does not consider failure factors;

[0031] 7) Each mission has only one aircraft to perform;

[0032] (3) Establish the reward function for distribution (Formula (2), (4)), considering:

[0033] 1) The aircraft cluster completes the mission as quickly as possible;

[0034] 2) The total fuel consumed by the aircraft cluster is the least;

[0035] Assume that the time taken for each rescue mission and delivery mission is T save and T send ;

[0036] Assume that the time taken for each rescue mission and delivery mission is T save and T send ;

[0037] Aircraft j needs to perform N save rescue missions and N send delivery tasks, where j = 1, 2..., N;

[0038] Total mission time T for aircraft j j for:

[0039]

[0040] Among them, T k-1:k T represents the time it takes for aircraft j to fly from the location where it performs mission (k-1) to the location where it performs mission k. 0:k represents the time it takes for aircraft j to travel from its starting point to the location where it performs mission k;

[0041] Based on the total mission time T of aircraft j jObtain the shortest time indicator for the entire aircraft cluster to complete the task;

[0042] Since the aircraft flies at a constant speed, fuel consumption is proportional to the distance traveled;

[0043] Assuming the unit fuel consumption is 1, the total mission distance D of aircraft j is j for:

[0044]

[0045] in,

[0046] D k-1:k represents the distance that aircraft j flies from the location where mission (k-1) is performed to the location where mission k is performed;

[0047] D 0:k represents the distance from the starting point to the location where task k is performed;

[0048] Based on the total mission distance D of aircraft j j Obtain the minimum fuel consumption index for the entire aircraft cluster to complete the mission;

[0049] (4) Based on the shortest time index for the entire aircraft cluster to complete the task and the minimum fuel consumption index for the entire aircraft cluster to complete the task, the task allocation problem is obtained;

[0050] (5) To describe the heterogeneity of tasks, the capability of aircraft j to perform different tasks is quantified by a number between [0,1], j = 1, 2..., N;

[0051] For example, if an aircraft carries detection and medical equipment but has little cargo space, its rescue capability can be quantified as 0.9 and its delivery capability can be quantified as 0.05. If an aircraft has a large interior space but no rescue equipment, its rescue capability can be quantified as 0.1 and its delivery capability can be quantified as 0.9. If an aircraft has both rescue equipment and interior space, its rescue capability can be quantified as 0.6 and its delivery capability can be quantified as 0.5.

[0052] Thus, the aircraft performance matrix X is established;

[0053] At the same time, set the lower limit of task capability requirements and establish the capability lower limit matrix Y;

[0054] (6) Based on the aircraft performance matrix X and the capability lower limit matrix Y, the constraints of mission planning are constructed;

[0055] Thus, the integer programming problem model with respect to time, fuel index and capacity constraints has been established (Equation (1)-Equation (8));

[0056] Other steps and parameters are the same as those in the first embodiment.

[0057] Specific implementation method three: This implementation method is different from specific implementation methods one or two in that: the total mission time T based on aircraft j in (3) j Obtain the shortest time indicator for the entire aircraft cluster to complete the task; the specific process is:

[0058] The shortest time indicator for the entire aircraft cluster to complete the mission is:

[0059] min J time =max{T j |j=1,2...,N} (2)

[0060] Among them, J time Represents the total mission time of the entire aircraft cluster.

[0061] Other steps and parameters are the same as those in the first or second embodiment.

[0062] Specific embodiment 4: This embodiment differs from any one of the specific embodiments 1 to 3 in that: the total mission distance D based on the aircraft j in (3) j Obtain the minimum fuel consumption index for the entire aircraft cluster to complete the mission; the specific process is:

[0063] The minimum fuel consumption index for the entire aircraft cluster to complete the mission is:

[0064]

[0065] Among them, J distance Represents the total mission distance of the entire aircraft cluster.

[0066] The other steps and parameters are the same as those in the first to third embodiments.

[0067] Specific embodiment 5: This embodiment differs from any one of specific embodiments 1 to 4 in that: in (4), the task allocation problem is obtained based on the shortest time index for the entire aircraft cluster to complete the task and the minimum fuel consumption index for the entire aircraft cluster to complete the task; it is expressed as follows:

[0068] min J=aJ time +bJ distance

[0069]

[0070] in,

[0071] J represents the task allocation function;

[0072] a and b are weighted coefficients for balancing time and distance dimensions, respectively, satisfying a + b = 1;

[0073] u ij Indicates whether the jth aircraft performs the i-th mission. If it does, u ij =1, otherwise u ij =0;

[0074] u represents the discriminant matrix of whether the aircraft performs the mission.

[0075] Other steps and parameters are the same as those in Specific Embodiments 1 to 4-1.

[0076] Specific embodiment 6: This embodiment differs from any one of specific embodiments 1 to 5 in that: in (5), the aircraft performance matrix X is established:

[0077]

[0078] in,

[0079] The first column represents the capability value of each aircraft to perform rescue missions;

[0080] The second column indicates the capability value of each aircraft to perform the delivery mission;

[0081] x 11 Indicates the ability value of the first aircraft to perform rescue mission, x 21 Indicates the ability value of the second aircraft to perform rescue missions, x N1 Indicates the capability value of the Nth aircraft to perform rescue missions;

[0082] x 12 Indicates the capability value of the first aircraft to perform the delivery mission, x 22 Indicates the capability value of the second aircraft to perform the delivery mission, x N2 Indicates the capability value of the Nth aircraft to perform the delivery mission;

[0083] At the same time, set the lower limit of task capability requirements and establish the capability lower limit matrix Y:

[0084]

[0085] Among them, y i Represents the lower limit of the required capacity of the i-th task, i = 1, 2,…, M.

[0086] Different rescue missions or delivery missions have different difficulties and details, so the capability requirement is the lower limit value y i different.

[0087] Other steps and parameters are the same as those in Specific Implementations 1 to 5-1.

[0088] Specific embodiment 7: This embodiment differs from any one of specific embodiments 1 to 6 in that: in (6), the constraints of the mission planning are constructed based on the aircraft performance matrix X and the capability lower limit matrix Y; the specific process is:

[0089] The constraints of task planning are expressed as:

[0090]

[0091] u ij =1 means the jth aircraft performs the ith mission, u ij =1, the capability value of aircraft j to perform rescue mission or delivery mission It must be higher than the lower limit y of the required capacity of the i-th task i ;

[0092] Represents x j1 or x j2 ; j=1,2...,N.

[0093] The other steps and parameters are the same as those in the first to sixth embodiments.

[0094] Specific embodiment eight: This embodiment differs from any one of specific embodiments one to seven in that: in step two, an improved wolf pack optimization algorithm is used to optimize the optimal multi-task allocation model for heterogeneous aircraft to obtain the optimal task allocation solution; the specific process is as follows:

[0095] The task allocation problem includes indicators, variables, and constraints; that is, under the constraints of Equation 8, find u in Equation 5 to minimize J in Equation 5.

[0096] Wolves rely on teamwork to hunt for food and are divided into alpha wolves, scout wolves, and fierce wolves according to their individual responsibilities.

[0097] There is competition within the wolf pack and inferior individuals are gradually eliminated. The main mechanisms are as follows:

[0098] 1) Alpha wolf generation: The individual with the best objective function value is randomly selected from among the individuals if there is a tie for first place. The alpha wolf is reselected in each iteration or replaced by another individual during the hunting process;

[0099] 2) Wandering mechanism: Except for the leader, the Q with better objective function t The wolf is a scouting wolf, with a walking step length of S t Wandering, including Q t is a random integer;

[0100] Walking step length S t , Summoning step length S m , siege step length S w It is a relative value, such as St =2, S w =1, S m =3;

[0101] If the target function value of the scout wolf exceeds that of the leader wolf during the wandering process, the scout wolf becomes the new leader wolf and performs the calling behavior, otherwise it continues to wander until the maximum number of wandering times is reached;

[0102] 3) Summoning behavior: After the wolf leader issues a command, Q m A wolf is summoned with a step length of S m Approaching the alpha wolf, if the objective function value of the fierce wolf exceeds that of the alpha wolf during the approach, the fierce wolf becomes the new alpha wolf and performs the siege behavior, otherwise it continues to approach until the summoning distance threshold is reached;

[0103] 4) Siege behavior: Exploring wolves and fierce wolves attack with a siege step length S w Explore around the alpha wolf and update the positions of the exploratory and ferocious wolves based on the better objective function values ​​of the exploratory and ferocious wolves;

[0104] 5) Elimination mechanism: Individuals with poor objective function values ​​will be eliminated and then redistributed to the solution space to form a new search group to participate in the next hunting. The number of eliminations is a random integer R. The more individuals are eliminated, the more random the search algorithm tends to be and the slower the convergence. Otherwise, the wolf pack tends to be stable and easily falls into a local optimum.

[0105] like Figure 1 , based on the above description, an improved wolf pack algorithm optimization is proposed:

[0106] 1) Chaotic reverse learning initialization;

[0107] 2) Divide the wolf pack:

[0108] Calculate the task allocation function value of each wolf in the entire wolf pack and sort them from small to large;

[0109] The wolf with the smallest task allocation function value is the leader. If there is a tie for first place, a wolf is randomly selected as the leader.

[0110] Except for the alpha wolf, the task allocation function has a smaller value Q t The wolf individual is the wolf, Q t is a random positive integer;

[0111] Except for the alpha wolf and the scout wolf, all other wolves are fierce wolves;

[0112] The number of the leader wolf is greater than or equal to 1, the number of the scout wolf is greater than or equal to 1, and the number of the fierce wolf is greater than or equal to 1;

[0113] 3) Wandering behavior:

[0114] The alpha wolf issues a wandering command, and the scout wolf wanders with a walking step length S tWandering, during the wandering process of the scout wolf, determine whether the task allocation function value of the scout wolf is less than that of the leader wolf;

[0115] If the task allocation function value of the scout wolf is less than that of the alpha wolf during its wandering process, the scout wolf will replace the alpha wolf as the new alpha wolf and immediately stop wandering, and execute 4);

[0116] If the task allocation function value of the scout wolf is greater than or equal to the leader wolf during the scout wolf's wandering process, the scout wolf will continue to wander with a step length of S t Wander until the maximum number of wandering times is reached, and then execute 4);

[0117] Here we use the Mantegna algorithm to simulate Levy flight, forming a walking step length S t for:

[0118]

[0119] Among them, S t represents the wolf's wandering step length, β is a random number between [0,2], round represents rounding, μ and v are random numbers that obey Gaussian distribution, and the variances of μ and v are σ respectively. μ and σ v :

[0120]

[0121] σ v =1

[0122] Where γ represents the gamma distribution;

[0123] 4) Summoning behavior:

[0124] The alpha wolf issues a summoning command, and the fierce wolf takes a summoning step length S m Charge towards the alpha wolf. During the charge, determine whether the task allocation function value of the alpha wolf is smaller than that of the alpha wolf.

[0125] If the task allocation function value of the fierce wolf is smaller than that of the alpha wolf during the process of the fierce wolf charging towards the alpha wolf, the fierce wolf will replace the alpha wolf as the new alpha wolf and immediately stop the summoning behavior, and execute 5);

[0126] If the wolf's task allocation function value is greater than or equal to the alpha wolf during the wolf's attack, the wolf will continue to summon the alpha wolf with a step length of S. m Run towards the leader wolf until the distance between them is less than the summoning distance threshold, then execute 5).

[0127] 5) Behavior of siege:

[0128] The alpha wolf issues a siege command, and the scout wolves and fierce wolves attack with a siege step length of S. wExplore around the alpha wolf, and update the positions of the exploratory and ferocious wolves based on the smaller value of the exploratory and ferocious wolves' task allocation functions. After the exploratory and ferocious wolves have updated their positions, execute step 6).

[0129] 6) Elimination behavior:

[0130] Calculate the task allocation function value of each wolf individual, sort the task allocation function values ​​from small to large, and eliminate the R wolf individuals at the bottom; R is a random integer;

[0131] 7) Repeat steps 2)-6) until the maximum number of iterations is reached, obtain the optimal task allocation function value, and obtain the optimal allocation plan based on the optimal task allocation function value.

[0132] The optimal allocation model forms the foundation upon which the optimization algorithm is deployed to solve and arrive at the optimal allocation solution. For example, the allocation model specifies the optimization variables and indicators. The optimization variables represent the state of the wolf pack in the optimization algorithm. Individual wolves correspond to indicators one-to-one, and the optimal solution for the entire pack is the maximization / minimization of the corresponding indicator. The pack continuously updates its state and gradually improves the indicator, a process known as optimization.

[0133] The task assignment problem includes rewards (or indicators) (the reward is the task assignment function J in Equation (5)), positions (the discriminant matrix u of whether the aircraft performs the task in Equation (5)), and constraints (Equation 8); that is, under the constraints of Equation 8, find u in Equation 5 so that J in Equation (5) is minimized.

[0134] Other steps and parameters are the same as those in Specific Embodiments 1 to 7-1.

[0135] Specific embodiment 9: This embodiment differs from specific embodiments 1 to 8 in that: 1) the chaotic reverse learning is initialized; the specific process is:

[0136] 1a) Generate a random sequence of tents:

[0137] For a scenario where N aircraft perform M tasks together, first a random number r1 between [0,1] is given, and the recursive method of the Tent random sequence is:

[0138]

[0139] Where n represents the random number, n = 1, 2, ..., Q / 2; ζ is a random number between [0, 0.1]; r n+1 Represents the n+1th random number; r n Represents the nth random number;

[0140] 1b), population initialization;

[0141] 1c) Check constraints;

[0142] 1d) Determine the parameters: maximum number of iterations, maximum number of walks, summoning distance threshold, wolf pack update ratio, summoning step length S m and siege step length S w .

[0143] The other steps and parameters are the same as those in the specific implementation modes 1 to 8-1.

[0144] Specific embodiment 10: This embodiment differs from any one of specific embodiments 1 to 9 in that: in the above 1b) the population is initialized:

[0145] 1b1) Considering the wolf pack size to be Q, an initial wolf pack consisting of Q / 2 wolf individuals is generated based on the Tent random sequence. Q / 2 wolf individuals correspond to Q / 2 allocation results.

[0146] The wolf pack includes the position and reward (or index) of each individual wolf. The position of each individual wolf is the discriminant matrix u of whether the aircraft performs the task in formula (5), and the reward is the task allocation function J in formula (5);

[0147] The Q / 2 allocation results include Q / 2 aircraft’s decision matrices u for determining whether to execute the mission and Q / 2 mission allocation functions J;

[0148] Q / 2 u correspond to u1,u2,…,u Q / 2 , u1 is the discriminant matrix of whether the first aircraft performs the mission, u2 is the discriminant matrix of whether the second aircraft performs the mission, u Q / 2 is the discriminant matrix of whether the Q / 2th aircraft performs the mission;

[0149] Q / 2 task allocation functions J correspond to J1, J2, ..., J Q / 2 , J1 is the first task allocation function, J2 is the second task allocation function, J Q / 2 Assign a function to the Q / 2th task;

[0150] 1b2) Using a reverse learning strategy, find an inverse solution for each wolf individual in the initial wolf pack consisting of Q / 2 wolf individuals generated in 1b1) to obtain Q / 2 inverse solutions, each of which corresponds to Q / 2 allocation results;

[0151] The Q / 2 allocation results include Q / 2 aircraft’s decision matrices u for determining whether to execute the mission and Q / 2 mission allocation functions J;

[0152] Q / 2 u correspond to u Q / 2+1 ,u Q / 2+2 ,…,u Q ,u Q / 2+1 is the discriminant matrix of whether the Q / 2+1th aircraft performs the mission, uQ / 2+2 is the discriminant matrix of whether the Q / 2+2th aircraft performs the mission, u Q is the discriminant matrix of whether the Qth aircraft performs the mission;

[0153] Q / 2 task allocation functions J correspond to J Q / 2+1 ,J Q / 2+2 ,…,J Q , J Q / 2+1 Assign function to the first task, J Q / 2+2 Assign function to the second task, J Q Assign a function to the Q / 2th task.

[0154] The specific process of checking constraints in 1c) is as follows:

[0155] Check the constraint relationship. If u1,u2,…,u Q If constraint (8) is satisfied, execute 2), otherwise execute 1a) to 1b) until u1,u2,…,u Q Satisfy constraint (8).

[0156] The other steps and parameters are the same as those in the specific implementation modes 1 to 9-1.

[0157] The present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.

Claims

1. A heterogeneous aircraft collaborative mission planning method based on improved wolf pack, characterized by: The specific process of the method is: Step 1: Construct an optimal task allocation model for heterogeneous aircraft; Step 2: Use the improved wolf pack optimization algorithm to optimize the optimal task allocation model of heterogeneous aircraft and obtain the optimal task allocation plan.

2. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 1, characterized in that: In the step 1, an optimal allocation model for heterogeneous aircraft tasks is constructed; the specific process is as follows: (1) Suppose N aircrafts jointly perform M tasks; The tasks are divided into rescue tasks and delivery tasks, and the types are represented by K i Indicates that K i =1 means rescue mission, K i =2 indicates the delivery task, i=1,2,…,M; (2) Make the following assumptions about the aircraft task allocation problem: 1) Terrain obstacles and unexpected threats are not considered; 2) Spatial distance is expressed by two-dimensional Euclidean distance; 3) The destination corresponding to each task remains stationary; 4) The aircraft flies at a constant speed; 5) The aircraft selects the task closest to the current position as the next task; 6) The aircraft has no communication distance and fuel restrictions, and does not consider failure factors; 7) Each mission has only one aircraft to perform; (3) Establish the reward function for distribution, considering: 1) The aircraft cluster completes the mission as quickly as possible; 2) The total fuel consumed by the aircraft cluster is minimal; Assume that the time taken for each rescue mission and delivery mission is T save and T send ; Aircraft j needs to perform N save rescue missions and N send delivery tasks, where j = 1, 2..., N; Total mission time T of aircraft j j for: Among them, T k-1:k T represents the time it takes for aircraft j to fly from the location of mission (k-1) to the location of mission k, 0:k represents the time it takes for aircraft j to travel from its starting point to the location where it performs mission k; Based on the total mission time T of aircraft j j Obtain the shortest time indicator for the entire aircraft cluster to complete the task; Assuming the unit fuel consumption is 1, the total mission distance D of aircraft j is j for: in, D k-1:k represents the distance that aircraft j flies from the location where mission (k-1) is performed to the location where mission k is performed; D 0:k represents the distance from the starting point to the location where task k is performed; Based on the total mission distance D of aircraft j j Obtain the minimum fuel consumption index for the entire aircraft cluster to complete the task; (4) Based on the shortest time index for the entire aircraft cluster to complete the task and the minimum fuel consumption index for the entire aircraft cluster to complete the task, the task allocation problem is obtained; (5) Quantify the capability of aircraft j to perform different tasks with a number between [0,1], j = 1, 2, ..., N; Establish aircraft performance matrix X; At the same time, set the lower limit of task capability requirements and establish the capability lower limit matrix Y; (6) Based on the aircraft performance matrix X and the capability lower limit matrix Y, construct the constraints for mission planning; At this point, an integer programming problem model with respect to time, fuel indicators, and capacity constraints has been established.

3. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 2 is characterized in that: The total mission time T based on aircraft j in (3) j Obtain the shortest time indicator for the entire aircraft cluster to complete the task; the specific process is: The shortest time indicator for the entire aircraft cluster to complete the task is: min J time =max{T j |j=1,2...,N} (2) Among them, J time Represents the total mission time of the entire aircraft cluster.

4. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 3 is characterized in that: The total mission distance D based on aircraft j in (3) j Obtain the minimum fuel consumption index for the entire aircraft cluster to complete the task; the specific process is: The minimum fuel consumption index for the entire aircraft cluster to complete the mission is: Among them, J distance Represents the total mission distance of the entire aircraft cluster.

5. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 4 is characterized in that: In (4), the task allocation problem is obtained based on the shortest time index for the entire aircraft cluster to complete the task and the minimum fuel consumption index for the entire aircraft cluster to complete the task; it is expressed as follows: my J=aJ time +bJ distance in, J represents the task allocation function; a and b are weighted coefficients for balancing the time and distance dimensions, respectively, satisfying a+b=1; u ij Indicates whether the jth aircraft executes the i-th mission. If it does, u ij =1, otherwise u ij =0; u represents the discriminant matrix of whether the aircraft performs the mission.

6. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 5 is characterized in that: In (5), the aircraft performance matrix X is established: in, The first column indicates the capability value of each aircraft to perform rescue missions; The second column indicates the capability value of each aircraft to perform the delivery mission; x 11 represents the capability value of the first aircraft to perform the rescue mission, x 21 Indicates the ability of the second aircraft to perform rescue missions, x N1 Indicates the capability value of the Nth aircraft to perform the rescue mission; x 12 Indicates the capability value of the first aircraft to perform the delivery mission, x 22 Indicates the capability value of the second aircraft to perform the delivery mission, x N2 Indicates the capability value of the Nth aircraft to perform the delivery mission; At the same time, set the lower limit of task capability requirements and establish the capability lower limit matrix Y: Among them, y i Represents the lower limit of the required capacity of the i-th task, i=1,2,…,M.

7. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 6 is characterized in that: In (6), based on the aircraft performance matrix X and the capability lower limit matrix Y, the constraints of the mission planning are constructed; the specific process is: The constraints of task planning are expressed as: u ij =1 means the jth aircraft performs the ith mission, u ij = 1, the capability value of aircraft j to perform rescue mission or delivery mission It must be higher than the lower limit y of the required capacity of the i-th task i ; Represents x j1 or x j2 ; j=1,2...,N.

8. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 7 is characterized in that: In the step 2, the improved wolf pack optimization algorithm is used to optimize the optimal task allocation model of heterogeneous aircraft to obtain the optimal task allocation solution; the specific process is: 1) Chaotic reverse learning initialization; 2) Divide the wolf pack: Calculate the task allocation function value of each individual wolf in the entire wolf pack and sort them from small to large; The wolf with the smallest task allocation function value is the leader. If there is a tie for first place, a wolf is randomly selected as the leader. Except for the alpha wolf, the task allocation function has a smaller value Q t The wolf individual is a stalker wolf, Q t is a random positive integer; Except for the alpha wolf and the scout wolf, all other wolves are fierce wolves; 3) Wandering behavior: The alpha wolf issues a wandering command, and the scout wolf wanders with a walking step length S t Wandering, during the wandering process of the scout wolf, determine whether the task allocation function value of the scout wolf is less than that of the leader wolf; If the task allocation function value of the scout wolf is less than that of the alpha wolf during its wandering, the scout wolf will replace the alpha wolf as the new alpha wolf and immediately stop wandering, and execute 4); If the task allocation function value of the scout wolf is greater than or equal to the leader wolf during the scout wolf's wandering process, the scout wolf continues to wander with a step length S t Wander until the maximum number of wandering times is reached, and then execute 4); Walking step length S t for: Among them, S t represents the walking step length of the wolf, β is a random number between [0,2], round means rounding to the nearest integer, μ and v are random numbers that follow a Gaussian distribution, and the variances of μ and v are σ respectively. μ and σ v : s v =1 Where γ represents the gamma distribution; 4) Summoning behavior: The alpha wolf issues a summoning command, and the fierce wolf takes a summoning step length S. m Run towards the leader wolf, and during the run, determine whether the task allocation function value of the leader wolf is smaller than that of the leader wolf; If the task allocation function value of the fierce wolf is smaller than that of the alpha wolf during the process of the fierce wolf rushing towards the alpha wolf, the fierce wolf replaces the alpha wolf as the new alpha wolf and immediately stops the summoning behavior, and executes 5); If the wolf's task allocation function value is greater than or equal to the alpha wolf during the wolf's charge, the wolf continues to summon with a step length S. m Run towards the leader wolf until the distance between the fierce wolf and the leader wolf is less than the summoning distance threshold, and then execute 5); 5) Behavior of siege: The leader wolf issues a siege command, and the scout wolves and fierce wolves attack with a siege step length of S. w Explore around the alpha wolf, and update the positions of the exploratory wolf and the fierce wolf based on the smaller value of the exploratory wolf and the fierce wolf's task allocation function. After the exploratory wolf and the fierce wolf update their positions, execute 6); 6) Elimination behavior: Calculate the task allocation function value of each wolf individual, sort the task allocation function values ​​from small to large, and eliminate the R wolf individuals at the bottom; R is a random integer; 7) Repeat steps 2)-6) until the maximum number of iterations is reached, obtain the optimal task allocation function value, and obtain the optimal allocation plan based on the optimal task allocation function value.

9. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 8, characterized in that: The chaotic reverse learning initialization in 1) is as follows: 1a) Generate a random sequence of tents: For a scenario where N aircraft perform M tasks together, first a random number r1 between [0,1] is given, then the recursive method of the Tent random sequence is: Where n represents the number of the random number, n = 1, 2, ..., Q / 2; ζ is a random number between [0, 0.1]; r n+1 Represents the n+1th random number; r n Represents the nth random number; 1b), population initialization; 1c) Check constraints; 1d) Determine the parameters: maximum number of iterations, maximum number of walks, summoning distance threshold, wolf pack update ratio, summoning step length S m and siege step length S w .

10. The method for collaborative mission planning of heterogeneous aircraft based on improved wolf pack according to claim 9, characterized in that: The population initialization in 1b) is: 1b1) Considering the wolf pack size as Q, an initial wolf pack consisting of Q / 2 wolf individuals is generated based on the Tent random sequence, and Q / 2 wolf individuals correspond to Q / 2 allocation results; The wolf pack includes the position and reward of each individual wolf. The position of each individual wolf is the discriminant matrix u of whether the aircraft performs the task in formula (5), and the reward is the task allocation function J in formula (5); Q / 2 allocation results include Q / 2 aircraft execution task determination matrices u and Q / 2 task allocation functions J; Q / 2 u correspond to u1,u2,…,u Q / 2 , u1 is the discriminant matrix of whether the first aircraft performs the mission, u2 is the discriminant matrix of whether the second aircraft performs the mission, u Q / 2 is the discriminant matrix of whether the Q / 2th aircraft performs the mission; Q / 2 task allocation functions J correspond to J1, J2, …, J Q / 2 , J1 is the first task allocation function, J2 is the second task allocation function, J Q / 2 Assign a function to the Q / 2th task; 1b2), using the reverse learning strategy to find the inverse solution for each wolf individual in the initial wolf pack consisting of Q / 2 wolf individuals generated in 1b1), and obtain Q / 2 inverse solutions, and Q / 2 inverse solutions correspond to Q / 2 allocation results; Q / 2 allocation results include Q / 2 aircraft execution task determination matrices u and Q / 2 task allocation functions J; Q / 2 u correspond to u Q / 2+1 ,u Q / 2+2 ,…,u Q ,u Q / 2+1 is the discriminant matrix of whether the Q / 2+1th aircraft performs the mission, u Q / 2+2 is the discriminant matrix of whether the Q / 2+2th aircraft performs the mission, u Q is the discriminant matrix of whether the Qth aircraft performs the mission; Q / 2 task allocation functions J correspond to J Q / 2+1 ,J Q / 2+2 ,…,J Q , J Q / 2+1 Assign function to the first task, J Q / 2+2 Assign a function to the second task, J Q Assign a function to the Q / 2th task. The check constraint in 1c) is as follows: Check the constraint relationship. If u1,u2,…,u Q If constraint (8) is satisfied, execute 2), otherwise execute 1a) to 1b) until u1,u2,…,u Q Satisfy constraint (8).

Images