Multi-UAV task allocation system and method based on dynamic wolf pack algorithm

Abstract

The present invention provides a multi-UAV task allocation system based on a dynamic wolf pack algorithm, including: an initialization module, used to set the position of each individual wolf in the wolf pack to a multi-UAV task allocation scheme composed of a task number list and a UAV number list, and the task number corresponds to the UAV number one by one; a wandering behavior module, used to use an ordered arrangement method to calculate the maximum wandering times of each exploratory wolf in the individual wolf and then wander; a summoning behavior module, used to make the fierce wolf summon under the premise that the task number list remains unchanged; a siege behavior module, used to calculate the corresponding distance probability according to the fitness difference between each individual wolf except the current head wolf and the current head wolf in the individual wolf, and then siege; an individual wolf update behavior module, used to update the individual wolf under the premise of ensuring that the task number list remains unchanged. The present invention is improved based on the wolf pack algorithm to obtain a dynamic wolf pack algorithm that is more suitable for multi-UAV task allocation.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-UAV task allocation, and in particular to a multi-UAV task allocation system and method based on a dynamic wolf pack algorithm. Background Art

[0002] In recent years, drones have attracted extensive attention at home and abroad due to their low cost, high mobility and wide applicability. Thanks to the continuous development of artificial intelligence and swarm control technology, and the fact that a single drone is limited by flight distance, load and other factors when performing a specific task, it is difficult to effectively perform the task alone, making multi-drone collaboration the focus of current research.

[0003] At present, swarm intelligence algorithms are mostly used to solve the problem of multi-UAV task allocation. Among swarm intelligence algorithms, particle swarm optimization has become the mainstream algorithm for multi-UAV task allocation due to its advantages of few parameters, simple algorithm implementation, convergence, and fast solution speed. Although particle swarm optimization has achieved good results in processing simple, unimodal, and continuous functions, it is prone to problems such as premature convergence, low solution accuracy, and falling into local optimal solutions when solving the complex discrete problem of multi-UAV task allocation.

[0004] As a newer swarm intelligence algorithm, the core idea of ​​the wolf pack algorithm is to imitate the hunting behavior of wolves. Compared with the particle swarm algorithm, the wolf pack algorithm has higher robustness and stronger global search capabilities, and also shows good performance in calculating high-dimensional and discrete problems. In recent years, wolf pack algorithms have achieved certain research results in the field of multi-UAV allocation. For example, the MDWPA algorithm proposes to combine the traditional wolf pack algorithm with the genetic algorithm and the elite strategy, which greatly improves the algorithm in terms of convergence speed and solution accuracy. However, the algorithm has defects such as low population diversity and easy selection of local optimality in the late stage of solution during the solution process; the CDWPA algorithm proposes to improve the quality of the initial solution through chaotic inverse initialization, and add random perturbations in the solution process to increase the global search ability of the artificial wolf pack, but the algorithm is prone to premature phenomenon in iteration, resulting in unstable solution effect; the MPPWPA algorithm proposes to divide the artificial wolf pack into multiple sub-groups, each sub-group solves independently, and realizes the continuous evolution of the sub-groups through information exchange between sub-groups, which greatly avoids the problem of decreased population diversity in the late stage of algorithm iteration under a single population. The algorithm shows good performance in solving low-dimensional problems, but with the increase of solution dimension, the accuracy of the algorithm solution decreases accordingly. Summary of the invention

[0005] The purpose of the present invention is to address the defects in the background technology and provide a multi-UAV task allocation system based on a dynamic wolf pack algorithm, comprising:

[0006] A multi-UAV task allocation system based on dynamic wolf pack algorithm, comprising:

[0007] An initialization module is used to define the position of each individual wolf in the wolf pack algorithm as a multi-UAV task allocation scheme consisting of a task number list and a UAV number list, wherein the task number in the task number list corresponds one-to-one to the UAV number in the UAV number list;

[0008] A wandering behavior module, used for calculating the maximum wandering times of each exploratory wolf in the individual wolf by using an ordered arrangement method, and wandering according to the maximum wandering times, wherein the wandering is defined as reordering the task numbers in the task number list;

[0009] A summoning behavior module, used to enable the fierce wolf among the individual wolves to summon the fierce wolf under the premise that the task number list remains unchanged after the task numbers are reordered, wherein the summoning is defined as the fierce wolf learning the multi-UAV task allocation plan of the current leader wolf;

[0010] A siege behavior module, which is used to calculate the corresponding distance probability according to the fitness difference between each individual wolf except the current alpha wolf and the current alpha wolf among the individual wolves, and to carry out siege according to the distance probability. The siege is defined as generating a random number within a preset range. If the distance probability is greater than the random number, the individual wolf mutates. If the distance probability is less than the random number, the individual wolf learns the multi-UAV task allocation scheme from the individual wolf with a smaller fitness than the individual wolf.

[0011] The individual wolf update behavior module is used to update the positions of individual wolves except the current alpha wolf after the above-mentioned wandering, summoning and siege, and under the premise of ensuring that the task number list remains unchanged, thereby forming new positions of all individual wolves, and outputting the current alpha wolf position among all new individual wolf positions. The position update of the individual wolf is defined as the L individual wolves ranked in ascending order of fitness value learning the multi-UAV task allocation plan of the top X individual wolves ranked in ascending order of fitness value.

[0012] Furthermore, the iterative calculation module is used to perform the wandering behavior module, the summoning behavior module, the siege behavior module and the individual wolf update behavior module in sequence as one round of iteration, set the maximum number of iterations to complete the iterative calculation, output the alpha wolf position generated by the last iteration, and use the alpha wolf position as the final multi-UAV task allocation plan.

[0013] Furthermore, the initialization module is also used to initialize the parameters of the dynamic wolf pack algorithm, including:

[0014] Population size, wolf detection scale factor, update scale factor, and temperature coefficient in the dynamic wolf pack algorithm.

[0015] Furthermore, in the wandering behavior module, the specific method of using the ordered arrangement method to calculate the maximum wandering times of each exploratory wolf in the individual wolf and wandering according to the maximum wandering times is:

[0016] First, the wolves are sorted in ascending order according to their fitness, and then an integer is randomly generated as the number of wolves to be explored in this iteration according to the formula for the number of wolves to be explored. The formula for the number of wolves to be explored is as follows:

[0017]

[0018] Among them, N is the population size, a is the wolf detection ratio factor;

[0019] Then, according to the task number list and the drone number list, the ordered arrangement method is used to calculate the maximum number of wandering times possnum for each scout wolf. The formula is as follows:

[0020]

[0021] in, is the total number of tasks for the mth UAV in the multi-UAV task allocation scheme for the i-th wolf scout;

[0022] Finally, the fitness of the scout wolf after wandering is calculated. If the fitness is lower than before wandering, the position of the scout wolf is updated. Otherwise, the scout wolf continues to wander until the maximum number of wandering times is reached.

[0023] Furthermore, in the summoning behavior module, the specific method for summoning the fierce wolf under the premise that the task number list remains unchanged is:

[0024] First, calculate the number of wolves fnum, the calculation formula is as follows:

[0025] fnum=N-snum-1

[0026] Among them, N is the population size, snum is the number of scouting wolves;

[0027] Then, keep the task number list of the wolf unchanged, randomly select j drone numbers in the drone number list of the wolf as the drone numbers to be modified, and modify the drone numbers to be modified to the drone numbers corresponding to the task numbers corresponding to the drone numbers to be modified in the drone number list of the alpha wolf, 1≤j≤the total number of numbers in the drone number list, and j is a randomly generated integer;

[0028] Finally, calculate the fitness of the wolf after summoning. If the fitness of the wolf after summoning is lower than that of the alpha wolf, the wolf will replace the alpha wolf as the new alpha wolf and initiate the summoning behavior again according to the new alpha wolf position. If the fitness of the wolf after summoning is lower than that before summoning but higher than that of the alpha wolf, the wolf will update its position. If the fitness of the wolf after summoning is higher than that before summoning, the position of the wolf will remain unchanged.

[0029] Furthermore, in the siege behavior module, the specific method of calculating the corresponding distance probability according to the fitness difference between each individual wolf except the current alpha wolf and the current alpha wolf among the individual wolves, and performing siege according to the distance probability is:

[0030] First, the calculation formula of distance probability P is as follows:

[0031]

[0032] Among them, F leader is the fitness of the alpha wolf, F i is the fitness corresponding to the i-th individual wolf, T is the temperature coefficient, and θ is the adjustment coefficient;

[0033] Calculate the probability P of the distance between the ith individual wolf and the leader wolf, and randomly generate a random number r between [0, 1]. If P>r, randomly select two drone numbers in the drone number list of the individual wolf for mutation; if P<r, randomly select the h pair of the individual wolf with corresponding task number and drone number to replace with: randomly select any h pair of the individual wolf with smaller fitness than the individual wolf with corresponding task number and drone number;

[0034] The mutation is defined as randomly selecting one among all drone numbers to replace the drone number to be mutated;

[0035] Finally, if the fitness of an individual wolf after the siege is lower than that before the siege, the position of the individual wolf is updated; otherwise, the position of the individual wolf remains unchanged.

[0036] Furthermore, in the individual wolf update behavior module, the specific method for updating the individual wolf under the premise of ensuring that the task number list remains unchanged is:

[0037] First, an integer is randomly generated as the number L of individual wolf updates in this round of iteration according to the update formula. The update formula is as follows:

[0038]

[0039] Among them, N is the population size, b is the update scale factor;

[0040] Then, under the premise of keeping the task number list unchanged, the drone number list of the L individual wolves ranked in ascending order of fitness values ​​is cleared, and at the same time, in the drone number list of the top X individual wolves in the wolf pack with ascending fitness values, the drone number with the most corresponding tasks is found as the preferred drone number, and the preferred drone number is filled into the drone number list of the L individual wolves ranked in ascending order of fitness values ​​according to the corresponding relationship between the preferred drone number and the task number.

[0041] Furthermore, the fitness calculation formula is as follows:

[0042] fitness=α×TCT+β×AFT+γ×ATF

[0043] Among them, fitness is the fitness, α is the weight of the time for the drone swarm to perform the reconnaissance process, β is the weight of the average flight time of the drone, γ is the weight of the average threat value of the drone, TCT is the time for the drone swarm to perform the reconnaissance process, AFT is the average flight time of each drone in the drone swarm, ATF is the average threat value faced by each drone in the drone swarm during the reconnaissance process, and the values ​​of α and β vary within the constraint of α+β=1(α, β∈(0,1)}. When the multi-drone task allocation scheme does not meet the constraint conditions, the fitness of the scheme is set to infinity. The reconnaissance process is a process from the starting point to the completion of all tasks and then returning to the starting point.

[0044] The calculation formula of the time TCT for the drone group to perform the reconnaissance process is as follows:

[0045]

[0046] In the formula, represents the time for the mth UAV to perform the reconnaissance process, Nu represents the number of UAVs, The calculation formula is as follows:

[0047]

[0048] In the formula, u m represents the mth drone, t k represents the kth task, represents the distance from the mth drone to the location of the first mission t1, represents the distance between the locations of the kth task and the k+1th task, x k represents the location of the kth task, represents the flight distance of the mth UAV to complete the kth mission, Indicates that the mth drone has the last mission t n The distance from the location to the return starting point, represents the flight speed of the mth UAV;

[0049] The calculation formula for the average flight time AFT of each drone in the drone group is as follows:

[0050]

[0051] Where, T m It represents the time required for the mth UAV to complete the reconnaissance process;

[0052] The calculation formula of the average threat value ATF faced by each drone in the drone group during reconnaissance is as follows:

[0053]

[0054] In the formula, It represents the threat coefficient faced by the m-th UAV when performing the k-th mission.

[0055] A multi-UAV task allocation method based on dynamic wolf pack algorithm, comprising:

[0056] The position of each individual wolf in the dynamic wolf pack algorithm is defined as a multi-UAV task allocation scheme consisting of a task number list and a UAV number list, where the task number in the task number list corresponds one-to-one to the UAV number in the UAV number list;

[0057] The maximum number of wanderings of each exploratory wolf in the individual wolf is calculated by using an ordered arrangement method, and wandering is performed according to the maximum number of wanderings, wherein the wandering is defined as reordering the task numbers in the task number list;

[0058] The fierce wolf among the individual wolves is made to summon under the premise that the task number list remains unchanged after the task numbers are reordered, and the summoning is defined as the fierce wolf learning the multi-UAV task allocation plan of the current leader wolf;

[0059] Calculate the corresponding distance probability according to the fitness difference between each individual wolf except the current leader wolf and the current leader wolf among the individual wolves, and carry out siege according to the distance probability, wherein the siege is defined as generating a random number within a preset range, if the distance probability is greater than the random number, the individual wolf mutates, and if the distance probability is less than the random number, the individual wolf learns the multi-UAV task allocation scheme from the individual wolf with a smaller fitness than the individual wolf;

[0060] After the above-mentioned wandering, summoning and siege, and under the premise of ensuring that the task number list remains unchanged, the positions of individual wolves except the current head wolf are updated to form new positions of all individual wolves, and the current head wolf position among all new positions of individual wolves is output. The position update of the individual wolf is defined as the L individual wolves ranked in ascending order of fitness values ​​learning the multi-UAV task allocation plan of the top X individual wolves ranked in ascending order of fitness values.

[0061] A computer program product includes a computer program / instruction, which, when executed by a processor, implements the multi-UAV task allocation method based on the dynamic wolf pack algorithm.

[0062] The beneficial effects of the present invention are:

[0063] 1. The present invention improves the wandering behavior of the scouting wolf. According to the tasks assigned to each drone and combined with the solution of the ordered arrangement problem in mathematical permutations and combinations, the maximum wandering times of each scouting wolf are calculated, and the maximum wandering times required by each scouting wolf are accurately matched.

[0064] 2. The present invention improves the calling behavior. Compared with the traditional calling behavior, since individual wolves use integer numbers and cannot use the distance formula for calculation, a genetic algorithm is combined, that is, the fierce wolf randomly selects several drone numbers in the drone number list, learns the multi-drone task allocation plan of the alpha wolf to update its own drone number list, and ensures the diversity of the population.

[0065] 3. The present invention improves the siege behavior. When an individual wolf is close to the alpha wolf, it mutates to ensure population diversity. When it is far away from the alpha wolf, it learns from wolves with lower fitness to make itself better. In the process of approaching the alpha wolf, it searches for the optimal solution and expands the solution space.

[0066] 4. The present invention improves the behavior of generating new individuals. The generation of new individuals is based on the multi-UAV task allocation plan of the individual wolves with the highest fitness values ​​in the wolf pack. The individual wolves are reconstructed instead of randomly generating new individuals. This ensures the quality of the regenerated individuals to a great extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a system block diagram of the present invention.

[0068] Figure 2 This is a flow chart of the dynamic wolf pack algorithm of the present invention.

[0069] Figure 3 Schematic diagram of the wandering behavior of the present invention.

[0070] Figure 4 It is a schematic diagram of the summoning behavior of the present invention.

[0071] Figure 5 It is a schematic diagram of the siege behavior of the present invention.

[0072] Figure 6 Generate a schematic diagram for a new individual of the present invention. DETAILED DESCRIPTION

[0073] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0074] Example 1

[0075] like Figure 1 As shown, a multi-UAV task allocation system based on a dynamic wolf pack algorithm includes:

[0076] An initialization module is used to define the position of each individual wolf in the wolf pack algorithm as a multi-UAV task allocation scheme consisting of a task number list and a UAV number list, wherein the task number in the task number list corresponds one-to-one to the UAV number in the UAV number list;

[0077] A wandering behavior module is used to calculate the maximum wandering times of each scout wolf in the individual wolf by using an ordered arrangement method, and wander according to the maximum wandering times to adjust the position of the scout wolf, wherein the wandering is defined as reordering the task numbers in the task number list;

[0078] A summoning behavior module is used to enable fierce wolves among individual wolves to summon the fierce wolves under the premise that the task number list remains unchanged after the task numbers are reordered, and adjust the positions of the fierce wolves. The summoning is defined as the fierce wolves learning the multi-UAV task allocation plan of the current leader wolf;

[0079] A siege behavior module is used to calculate the corresponding distance probability (the distance probability between the individual wolf and the current alpha wolf) according to the fitness difference between each individual wolf except the current alpha wolf and the current alpha wolf among the individual wolves, and to siege according to the distance probability, and adjust the position of each individual wolf except the current alpha wolf. The siege is defined as generating a random number within a preset range. If the distance probability is greater than the random number, the individual wolf mutates. If the distance probability is less than the random number, the individual wolf learns the multi-UAV task allocation scheme from the individual wolf with a smaller fitness than the individual wolf.

[0080] The individual wolf update behavior module is used to update the positions of individual wolves except the current alpha wolf after the above-mentioned wandering, summoning and siege, and under the premise of ensuring that the task number list remains unchanged, so as to form new positions of all individual wolves, and output the current alpha wolf position among all new positions of individual wolves. The position update of the individual wolf is defined as the L individual wolves ranked in ascending order of fitness value learning the multi-UAV task allocation scheme of the top X individual wolves ranked in ascending order of fitness value;

[0081] The iterative calculation module is used to perform the wandering behavior module, the summoning behavior module, the siege behavior module and the individual wolf update behavior module in sequence as one round of iteration, set the maximum number of iterations to complete the iterative calculation, output the position of the alpha wolf generated by the last iteration, and use the alpha wolf position as the final multi-UAV task allocation plan.

[0082] The fitness mentioned above, the smaller the fitness value, the better the fitness, and the larger the fitness value, the worse the fitness.

[0083] A new leader may emerge when the position of each individual wolf changes.

[0084] (1) The initialization module is also used to initialize the parameters of the dynamic wolf pack algorithm, including:

[0085] Population size, wolf detection scale factor, update scale factor, and temperature coefficient in the dynamic wolf pack algorithm.

[0086] (2) The specific method of the wandering behavior is:

[0087] First, the wolves are sorted in ascending order according to their fitness, and then the number of wolves is determined according to the formula for the number of wolves to be explored. A random integer is generated within the range as the number of wolf scouts in this iteration. The formula for the number of wolf scouts is as follows:

[0088]

[0089] Among them, N is the population size, a is the wolf detection ratio factor;

[0090] Then, according to the task number list and the drone number list, the ordered arrangement method is used to calculate the maximum number of wandering times possnum for each scout wolf. The formula is as follows:

[0091]

[0092] in, The total number of tasks for the mth UAV in the multi-UAV task allocation scheme for the i-th wolf explorer. When the maximum number of wandering times is too large, the maximum value truncation can be performed;

[0093] Finally, the fitness of the scout wolf after wandering is calculated. If the fitness is lower than before wandering, the scout wolf position is updated. Otherwise, the wandering behavior continues until the maximum number of wandering times is reached. The wandering behavior ends after all scout wolves have been traversed.

[0094] Each walk randomly disrupts the order in which the drone performs its tasks, such as Figure 3 As shown. The wandering behavior is completed by the scout wolf in the wolf pack. In the multi-UAV allocation problem, the wandering behavior can be understood as optimizing the execution order of the tasks assigned by the UAV so that the distance to execute the tasks is the shortest. Wandering once can be understood as sorting the execution order once. The wandering behavior of the traditional wolf pack algorithm needs to determine a fixed value as the maximum wandering number of all scout wolves. Since the number of tasks assigned to each UAV is different, the order that needs to be optimized is also different. Therefore, the present invention calculates the maximum wandering number of each scout wolf as the maximum wandering number in each wandering direction according to the tasks assigned to each UAV and the solution of the ordered arrangement problem in mathematical permutations and combinations to replace the fixed value of the traditional wolf pack algorithm. If the wandering number of the traditional algorithm is fixed to 10 times, if a UAV is only assigned one task, then there is no need to wander. If a UAV is assigned ten tasks, it is not enough to wander only 10 times. If a UAV is assigned three tasks, it only needs to wander 10 times to find the optimal task execution order, so the fixed wandering number either wastes computing power or the wandering number is not enough. The present invention calculates the maximum number of wandering times of each scouting wolf according to the tasks assigned to each drone and the solution of the ordered arrangement problem in mathematical permutations and combinations, and accurately matches the maximum number of wandering times required by each scouting wolf.

[0095] (3) The specific method of the summoning behavior is:

[0096] First, calculate the number of wolves fnum, the calculation formula is as follows:

[0097] fnum=N-snum-1

[0098] Among them, N is the population size, snum is the number of scouting wolves;

[0099] Then, keep the task number list of the wolf unchanged, randomly select j drone numbers in the wolf's drone number list as the drone numbers to be modified, and modify the drone numbers to be modified to the drone numbers corresponding to the task numbers corresponding to the drone numbers to be modified in the wolf's drone number list, 1≤j≤the total number of numbers in the drone number list, and j is a randomly generated integer. Figure 4As shown, for example, the fierce wolf selects drone numbers 1, 1, 2 as the drone numbers to be modified, and the corresponding task numbers are 2, 4, 6. Then the drone numbers 2, 1, 3 corresponding to the task numbers 2, 4, 6 in the wolf leader will replace the drone numbers 1, 1, 2 corresponding to the wolf's task numbers 2, 4, 6.

[0100] Finally, calculate the fitness of the wolf after summoning. If the fitness of the wolf after summoning is lower than that of the alpha wolf, then the wolf replaces the alpha wolf as the new alpha wolf and initiates the summoning behavior again according to the new alpha wolf position. If the fitness of the wolf after summoning is lower than before summoning but higher than that of the alpha wolf, then the wolf updates its position. If the fitness of the wolf after summoning is higher than before summoning, then the position of the wolf remains unchanged.

[0101] The summoning behavior is performed by the fierce wolf in the wolf pack, and the fierce wolf is an individual wolf with a larger fitness value. In fact, it is the process of the algorithm converging to the optimal solution. In the multi-UAV task allocation problem, the summoning behavior can be understood as a process in which the fierce wolf with a larger fitness value learns the multi-UAV task allocation scheme represented by the alpha wolf with the smallest fitness. Traditional wolf pack algorithms often use the position distance formula to update the position of the fierce wolf in the summoning behavior, but the present invention uses integer numbering and cannot use the distance formula for calculation, so it combines the genetic algorithm, that is, the fierce wolf randomly selects several UAV numbers in the UAV number list, learns the multi-UAV task allocation scheme of the alpha wolf to update its own UAV number list, and ensures the diversity of the population.

[0102] (4) The specific methods of the siege behavior are:

[0103] First, the calculation formula of distance probability P is as follows:

[0104]

[0105] Among them, F leader is the fitness of the alpha wolf, F i is the fitness corresponding to the i-th individual wolf, T is the temperature coefficient, and θ is the adjustment coefficient;

[0106] Calculate the probability P of the distance between the ith individual wolf and the leader wolf, and randomly generate a random number r between [0, 1]. If P>r, randomly select two drone numbers in the drone number list of the individual wolf for mutation; if P<r, randomly select the h pair of the individual wolf with corresponding task number and drone number to replace, and replace with: randomly select any h pair of the individual wolf with smaller fitness than the individual wolf with corresponding task number and drone number, such as Figure 5 As shown, two pairs of corresponding mission numbers and drone numbers are arbitrarily replaced.

[0107] The mutation is defined as randomly selecting one among all drone numbers to replace the drone number to be mutated;

[0108] If the fitness of the individual wolf after mutation is lower than that before mutation, the position of the individual wolf will be updated; otherwise, the position of the individual wolf will remain unchanged.

[0109] The traditional siege behavior sets a fixed distance threshold. When the position of an individual wolf and the alpha wolf is less than the distance threshold, the siege behavior is initiated. The improved siege behavior introduces the Metropolis criterion of the simulated annealing algorithm, and uses the fitness difference between the individual wolf and the alpha wolf to calculate the "distance probability". When the individual wolf is close to the alpha wolf, it mutates to ensure population diversity. When it is far away from the alpha wolf, it learns from wolves with lower fitness to make itself better, and searches for a better solution near its own position to expand the solution space range, instead of approaching the alpha wolf like a fixed distance threshold. It can achieve more self-mutation in the early stage of the individual, expand the search range, and accelerate convergence in the later stage. Secondly, if a fixed fitness is used instead of distance like the traditional wolf pack algorithm, it is not feasible in the entire iteration because the fitness difference between the alpha wolf and other individual wolves is greatly reduced in the later stage of iteration. Therefore, a dynamic method should be used to calculate the distance probability. In addition, the value of the temperature coefficient T will continue to decrease as the iteration proceeds. The lower the value of the temperature coefficient T, the lower the calculated distance probability. That is to say, even if the fitness difference between the alpha wolf and the individual wolves remains unchanged, the distance probability will continue to decrease as the iteration proceeds. The purpose is to make the individual wolves converge to the alpha wolf as the iteration proceeds. In the early stage of the iteration, the search range of the solution space can be expanded through mutation operations, and in the later stage, it is hoped to accelerate convergence.

[0110] (5) The specific method of updating the behavior of individual wolves is:

[0111] First, an integer is randomly generated as the number L of individual wolf updates in this round of iteration according to the update formula. The update formula is as follows:

[0112]

[0113] Among them, N is the population size, b is the update scale factor;

[0114] Then, under the premise of keeping the task number list unchanged, the drone number list of the L individual wolves ranked in ascending order of fitness values ​​is cleared, and at the same time, in the drone number list of the top X individual wolves in the wolf pack with the highest number of fitness values, the drone number with the most corresponding tasks is found as the preferred drone number, and the preferred drone number is filled into the drone number list of the L individual wolves ranked in ascending order of fitness values ​​according to the corresponding relationship between the preferred drone number and the task number, as shown in the following example: Figure 6As shown, for example, among the top 10 individual wolves in ascending order of fitness values, for Task 4, the drones that perform Task 4 among the 6 individual wolves are all drone 2, then drone number 2 is filled in the position corresponding to Task 4 in the drone number list of the next 20 individual wolves in ascending order of fitness values.

[0115] In the traditional wolf pack algorithm, individual regeneration is irregular. A batch of individual wolves are randomly regenerated in the solution space as the update of the wolf pack. However, the randomly generated individual wolves are difficult to compare with the individual wolves that are continuously optimized in the iterative process. Therefore, this algorithm improves the new individual generation mechanism, that is, the generation of new individuals is based on the multi-UAV task allocation plan of the top X individual wolves in the wolf pack in terms of fitness, and the individual wolves are reconstructed, which greatly guarantees the quality of the regenerated individuals. Generally, X can be set to 20%.

[0116] (6) The fitness calculation formula is as follows:

[0117] fitness=α×TCT+β×AFT+γ×ATF

[0118] Among them, fitness is the fitness, α is the weight of the time for the drone swarm to perform the reconnaissance process, β is the weight of the average flight time of the drone, γ is the weight of the average threat value of the drone, TCT is the time for the drone swarm to perform the reconnaissance process, AFT is the average flight time of each drone in the drone swarm, ATF is the average threat value faced by each drone in the drone swarm during the reconnaissance process, and the values ​​of α and β vary within the constraint condition of α+β=1{α, β∈(0,1)}. When the multi-drone task allocation scheme does not meet the constraint conditions, the fitness of the scheme is set to infinity. The reconnaissance process is a process from the starting point to the completion of all tasks and then returning to the starting point.

[0119] The time TCT of the drone performing the reconnaissance process is divided into the time required for the drone to fly to the target point, the time required for the drone to perform the reconnaissance process, and the time required for the drone to return to the launch platform. When all drones complete the reconnaissance process, it means that all reconnaissance processes are completed. The time TCT of the drone group performing the reconnaissance process is determined by the drone that takes the longest time. The calculation formula of the time TCT of the drone group performing the reconnaissance process is as follows:

[0120]

[0121] In the formula, represents the time for the mth UAV to perform the reconnaissance process, Nu represents the number of UAVs, The calculation formula is as follows:

[0122]

[0123] In the formula, u m represents the mth drone, t k represents the kth task, represents the distance from the mth drone to the location of the first mission t1, represents the distance between the locations of the kth task and the k+1th task, x k represents the location of the kth task, represents the flight distance of the mth UAV to complete the kth mission, Indicates that the mth drone has the last mission t n The distance from the location to the return starting point, represents the flight speed of the mth UAV;

[0124] The calculation formula for the average flight time AFT of each drone in the drone group is as follows:

[0125]

[0126] Where, T m It represents the time required for the mth UAV to complete the reconnaissance process;

[0127] The calculation formula of the average threat value ATF faced by each drone in the drone group during reconnaissance is as follows:

[0128]

[0129] In the formula, It represents the threat coefficient faced by the m-th UAV when performing the k-th mission.

[0130] Example 2

[0131] A multi-UAV task allocation method based on dynamic wolf pack algorithm, comprising:

[0132] The position of each individual wolf in the dynamic wolf pack algorithm is defined as a multi-UAV task allocation scheme consisting of a task number list and a UAV number list, where the task number in the task number list corresponds one-to-one to the UAV number in the UAV number list;

[0133] The maximum number of wanderings of each scout wolf in the individual wolf is calculated by using an ordered arrangement method, and wandering is performed according to the maximum number of wanderings to adjust the position of the scout wolf, wherein the wandering is defined as reordering the task numbers in the task number list;

[0134] The fierce wolves among the individual wolves are summoned on the premise that the task number list remains unchanged after the task numbers are reordered, and the positions of the fierce wolves are adjusted, wherein the summoning is defined as the fierce wolves learning the multi-UAV task allocation plan of the current leader wolf;

[0135] According to the fitness difference between each individual wolf except the current alpha wolf and the current alpha wolf among the individual wolves, the corresponding distance probability (the distance probability between the individual wolf and the current alpha wolf) is calculated respectively, and siege is carried out according to the distance probability, and the positions of each individual wolf except the current alpha wolf are adjusted. The siege is defined as generating a random number within a preset range. If the distance probability is greater than the random number, the individual wolf mutates. If the distance probability is less than the random number, the individual wolf learns the multi-UAV task allocation scheme from the individual wolf with a smaller fitness than the individual wolf.

[0136] After the above-mentioned wandering, summoning and siege, the positions of individual wolves except the current alpha wolf are updated while ensuring that the task number list remains unchanged, thereby forming new positions of all individual wolves, and outputting the current alpha wolf position among all new positions of individual wolves. The position update of the individual wolves is defined as the L individual wolves ranked in ascending order of fitness values ​​learning the multi-UAV task allocation scheme of the top X individual wolves ranked in ascending order of fitness values;

[0137] One round of iteration consists of performing the wandering behavior module, summoning behavior module, siege behavior module and individual wolf update behavior module in sequence. The maximum number of iterations is set to complete the iterative calculation, and the position of the alpha wolf generated by the last iteration is output, which is used as the final multi-UAV task allocation plan.

[0138] The fitness mentioned above, the smaller the fitness value, the better the fitness, and the larger the fitness value, the worse the fitness.

[0139] A new leader may emerge when the position of each individual wolf changes.

[0140] (1) Initialize the parameters of the dynamic wolf pack algorithm before wandering, including:

[0141] Population size, wolf detection scale factor, update scale factor, and temperature coefficient in the dynamic wolf pack algorithm.

[0142] (2) The specific method of the wandering behavior is:

[0143] First, the wolves are sorted in ascending order according to their fitness, and then the number of wolves is determined according to the formula for the number of wolves. A random integer is generated within the range as the number of wolf scouts in this iteration. The formula for the number of wolf scouts is as follows:

[0144]

[0145] Among them, N is the population size, a is the wolf detection ratio factor;

[0146] Then, according to the task number list and the drone number list, the ordered arrangement method is used to calculate the maximum number of wandering times possnum for each scout wolf. The formula is as follows:

[0147]

[0148] in, The total number of tasks for the mth UAV in the multi-UAV task allocation scheme for the i-th wolf explorer. When the maximum number of wandering times is too large, the maximum value truncation can be performed;

[0149] Finally, the fitness of the scout wolf after wandering is calculated. If the fitness is lower than before wandering, the scout wolf position is updated. Otherwise, the wandering behavior continues until the maximum number of wandering times is reached. The wandering behavior ends after all scout wolves have been traversed.

[0150] Each walk randomly disrupts the order in which the drone performs its tasks, such as Figure 3 As shown. The wandering behavior is completed by the scout wolf in the wolf pack. In the multi-UAV allocation problem, the wandering behavior can be understood as optimizing the execution order of the tasks assigned by the UAV so that the distance to execute the tasks is the shortest. Wandering once can be understood as sorting the execution order once. The wandering behavior of the traditional wolf pack algorithm needs to determine a fixed value as the maximum wandering number of all scout wolves. Since the number of tasks assigned to each UAV is different, the order that needs to be optimized is also different. Therefore, the present invention calculates the maximum wandering number of each scout wolf as the maximum wandering number in each wandering direction according to the tasks assigned to each UAV and the solution of the ordered arrangement problem in mathematical permutations and combinations to replace the fixed value of the traditional wolf pack algorithm. If the wandering number of the traditional algorithm is fixed to 10 times, if a UAV is only assigned one task, then there is no need to wander. If a UAV is assigned ten tasks, it is not enough to wander only 10 times. If a UAV is assigned three tasks, it only needs to wander 10 times to find the optimal task execution order, so the fixed wandering number either wastes computing power or the wandering number is not enough. The present invention calculates the maximum number of wandering times of each scouting wolf according to the tasks assigned to each drone and the solution of the ordered arrangement problem in mathematical permutations and combinations, and accurately matches the maximum number of wandering times required by each scouting wolf.

[0151] (3) The specific method of the summoning behavior is:

[0152] First, calculate the number of wolves fnum, the calculation formula is as follows:

[0153] fnum=N-snum-1

[0154] Among them, N is the population size, snum is the number of scouting wolves;

[0155] Then, keep the task number list of the wolf unchanged, randomly select j drone numbers in the wolf's drone number list as the drone numbers to be modified, and modify the drone numbers to be modified to the drone numbers corresponding to the task numbers corresponding to the drone numbers to be modified in the wolf's drone number list, 1≤j≤the total number of numbers in the drone number list, and j is a randomly generated integer. Figure 4 As shown, for example, the fierce wolf selects drone numbers 1, 1, 2 as the drone numbers to be modified, and the corresponding task numbers are 2, 4, 6. Then the drone numbers 2, 1, 3 corresponding to the task numbers 2, 4, 6 in the wolf leader will replace the drone numbers 1, 1, 2 corresponding to the wolf's task numbers 2, 4, 6.

[0156] Finally, calculate the fitness of the wolf after summoning. If the fitness of the wolf after summoning is lower than that of the alpha wolf, then the wolf replaces the alpha wolf as the new alpha wolf and initiates the summoning behavior again according to the new alpha wolf position. If the fitness of the wolf after summoning is lower than before summoning but higher than that of the alpha wolf, then the wolf updates its position. If the fitness of the wolf after summoning is higher than before summoning, then the position of the wolf remains unchanged.

[0157] The summoning behavior is performed by the fierce wolf in the wolf pack, and the fierce wolf is an individual wolf with a larger fitness value. In fact, it is the process of the algorithm converging to the optimal solution. In the multi-UAV task allocation problem, the summoning behavior can be understood as a process in which the fierce wolf with a larger fitness value learns the multi-UAV task allocation scheme represented by the alpha wolf with the smallest fitness. Traditional wolf pack algorithms often use the position distance formula to update the position of the fierce wolf in the summoning behavior, but the present invention uses integer numbering and cannot use the distance formula for calculation, so it combines the genetic algorithm, that is, the fierce wolf randomly selects several UAV numbers in the UAV number list, learns the multi-UAV task allocation scheme of the alpha wolf to update its own UAV number list, and ensures the diversity of the population.

[0158] (4) The specific methods of the siege behavior are:

[0159] First, the calculation formula of distance probability P is as follows:

[0160]

[0161] Among them, F leader is the fitness of the alpha wolf, F i is the fitness corresponding to the i-th individual wolf, T is the temperature coefficient, and θ is the adjustment coefficient;

[0162] Calculate the probability P of the distance between the ith individual wolf and the leader wolf, and randomly generate a random number r between [0, 1]. If P>r, randomly select two drone numbers in the drone number list of the individual wolf for mutation; if P<r, randomly select the h pair of the individual wolf with corresponding task number and drone number to replace, and replace with: randomly select any h pair of the individual wolf with smaller fitness than the individual wolf with corresponding task number and drone number, such as Figure 5 As shown, two pairs of corresponding mission numbers and drone numbers are arbitrarily replaced.

[0163] The mutation is defined as randomly selecting one among all drone numbers to replace the drone number to be mutated;

[0164] If the fitness of the individual wolf after mutation is lower than that before mutation, the position of the individual wolf will be updated; otherwise, the position of the individual wolf will remain unchanged.

[0165] The traditional siege behavior sets a fixed distance threshold. When the position of an individual wolf and the alpha wolf is less than the distance threshold, the siege behavior is initiated. The improved siege behavior introduces the Metropolis criterion of the simulated annealing algorithm, and uses the fitness difference between the individual wolf and the alpha wolf to calculate the "distance probability". When the individual wolf is close to the alpha wolf, it mutates to ensure population diversity. When it is far away from the alpha wolf, it learns from wolves with lower fitness to make itself better, and searches for a better solution near its own position to expand the solution space range, instead of approaching the alpha wolf like a fixed distance threshold. It can achieve more self-mutation in the early stage of the individual, expand the search range, and accelerate convergence in the later stage. Secondly, if a fixed fitness is used instead of distance like the traditional wolf pack algorithm, it is not feasible in the entire iteration because the fitness difference between the alpha wolf and other individual wolves is greatly reduced in the later stage of iteration. Therefore, a dynamic method should be used to calculate the distance probability. In addition, the value of the temperature coefficient T will continue to decrease as the iteration proceeds. The lower the value of the temperature coefficient T, the lower the calculated distance probability. That is to say, even if the fitness difference between the alpha wolf and the individual wolves remains unchanged, the distance probability will continue to decrease as the iteration proceeds. The purpose is to make the individual wolves converge to the alpha wolf as the iteration proceeds. In the early stage of the iteration, the search range of the solution space can be expanded through mutation operations, and in the later stage, it is hoped to accelerate convergence.

[0166] (5) The specific method of updating the behavior of individual wolves is:

[0167] First, an integer is randomly generated as the number L of individual wolf updates in this round of iteration according to the update formula. The update formula is as follows:

[0168]

[0169] Among them, N is the population size, b is the update scale factor;

[0170] Then, under the premise of keeping the task number list unchanged, the drone number list of the L individual wolves ranked in ascending order of fitness values ​​is cleared, and at the same time, in the drone number list of the top X individual wolves in the wolf pack with the highest number of fitness values, the drone number with the most corresponding tasks is found as the preferred drone number, and the preferred drone number is filled into the drone number list of the L individual wolves ranked in ascending order of fitness values ​​according to the corresponding relationship between the preferred drone number and the task number, as shown in the following example: Figure 6 As shown, for example, among the top 10 individual wolves in ascending order of fitness values, for Task 4, the drones that perform Task 4 among the 6 individual wolves are all drone 2, then drone number 2 is filled in the position corresponding to Task 4 in the drone number list of the next 20 individual wolves in ascending order of fitness values.

[0171] In the traditional wolf pack algorithm, individual regeneration is irregular. A batch of individual wolves are randomly regenerated in the solution space as the update of the wolf pack. However, the randomly generated individual wolves are difficult to compare with the individual wolves that are continuously optimized in the iterative process. Therefore, this algorithm improves the new individual generation mechanism, that is, the generation of new individuals is based on the multi-UAV task allocation plan of the top X individual wolves in the wolf pack in terms of fitness, and the individual wolves are reconstructed, which greatly guarantees the quality of the regenerated individuals. Generally, X can be set to 20%.

[0172] (6) The fitness calculation formula is as follows:

[0173] fitness=α×TCT+β×AFT+γ×ATF

[0174] Among them, fitness is the fitness, α is the weight of the time for the drone swarm to perform the reconnaissance process, β is the weight of the average flight time of the drone, γ is the weight of the average threat value of the drone, TCT is the time for the drone swarm to perform the reconnaissance process, AFT is the average flight time of each drone in the drone swarm, ATF is the average threat value faced by each drone in the drone swarm during the reconnaissance process, and the values ​​of α and β vary within the constraint condition of α+β=1{α, β∈(0,1)}. When the multi-drone task allocation scheme does not meet the constraint conditions, the fitness of the scheme is set to infinity. The reconnaissance process is a process from the starting point to the completion of all tasks and then returning to the starting point.

[0175] The time TCT of the drone performing the reconnaissance process is divided into the time required for the drone to fly to the target point, the time required for the drone to perform the reconnaissance process, and the time required for the drone to return to the launch platform. When all drones complete the reconnaissance process, it means that all reconnaissance processes are completed. The time TCT of the drone group performing the reconnaissance process is determined by the drone that takes the longest time. The calculation formula of the time TCT of the drone group performing the reconnaissance process is as follows:

[0176]

[0177] In the formula, represents the time for the mth UAV to perform the reconnaissance process, Nu represents the number of UAVs, The calculation formula is as follows:

[0178]

[0179] In the formula, u m represents the mth drone, t k represents the kth task, represents the distance from the mth drone to the location of the first mission t1, represents the distance between the locations of the kth task and the k+1th task, x k represents the location of the kth task, represents the flight distance of the mth UAV to complete the kth mission, Indicates that the mth drone has the last mission t n The distance from the location to the return starting point, represents the flight speed of the mth UAV;

[0180] The calculation formula for the average flight time AFT of each drone in the drone group is as follows:

[0181]

[0182] Where, T m It represents the time required for the mth UAV to complete the reconnaissance process;

[0183] The calculation formula of the average threat value ATF faced by each drone in the drone group during reconnaissance is as follows:

[0184]

[0185] In the formula, It represents the threat coefficient faced by the m-th UAV when performing the k-th mission.

[0186] Example 3

[0187] A computer program product includes a computer program / instruction, which, when executed by a processor, implements the multi-UAV task allocation method based on a dynamic wolf pack algorithm in Example 2.

[0188] The contents not described in detail in this specification belong to the technology known to professional and technical personnel in the field. It should be understood by those skilled in the art that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments combining software and hardware. Moreover, the present invention can adopt the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0189] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0190] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0191] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0192] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the pending claims of the invention.

Claims

1. A multi-UAV task allocation system based on dynamic wolf pack algorithm, characterized in that: Including: An initialization module, which is used to define the position of each individual wolf in the dynamic wolf pack algorithm as a multi-UAV task allocation scheme composed of a task number list and a UAV number list, where the task numbers in the task number list correspond one-to-one with the UAV numbers in the UAV number list; A wandering behavior module, which is used to calculate the maximum wandering times of each exploring wolf among the individual wolves by using the ordered arrangement method, and perform wandering according to the maximum wandering times, where the wandering is defined as reordering the task numbers in the task number list; A summoning behavior module, which is used to make the fierce wolves among the individual wolves perform summoning on the premise that the task number list remains unchanged after the task numbers are reordered, where the summoning is defined as the fierce wolves learning the multi-UAV task allocation scheme of the current leading wolf; A besieging behavior module, which is used to calculate the corresponding distance probabilities respectively according to the fitness differences between each individual wolf except the current leading wolf and the current leading wolf among the individual wolves, and perform besieging according to the distance probabilities, where the besieging is defined as generating a random number within a preset range. If the distance probability is greater than the random number, then this individual wolf mutates; if the distance probability is less than the random number, then this individual wolf learns the multi-UAV task allocation scheme from an individual wolf with a smaller fitness than this individual wolf. First, the calculation formula of the distance probability P is as follows: Among them, F leader is the fitness of the alpha wolf, F i is the fitness corresponding to the i-th individual wolf, T is the temperature coefficient, and θ is the adjustment coefficient; Calculate the distance probability P between the i-th individual wolf and the leading wolf, and at the same time randomly generate a random number r between [0, 1]. If P > r, then randomly select two UAV numbers in the UAV number list of this individual wolf for mutation; if P < r, then randomly select h pairs of corresponding task numbers and UAV numbers of this individual wolf for replacement, and the replacement is: randomly select any h pairs of corresponding task numbers and UAV numbers of an individual wolf with a smaller fitness than this individual wolf; The mutation is defined as randomly selecting one from all the UAV numbers to replace the UAV number to be mutated; Finally, if the fitness of the individual wolf after besieging is lower than that before besieging, then update the position of this individual wolf; otherwise, the position of this individual wolf remains unchanged; An individual wolf update behavior module, which is used to update the positions of the individual wolves except the current leading wolf after performing the above-mentioned wandering, summoning and besieging, and on the premise of ensuring that the task number list remains unchanged, so as to form new positions of all individual wolves, and output the position of the current leading wolf in the new positions of all individual wolves. The position update of the individual wolf is defined as that the L individual wolves with ascending fitness values learn the multi-UAV task allocation scheme of the top X individual wolves with ascending fitness values.

2. The multi-UAV task allocation system based on dynamic wolf pack algorithm according to claim 1 is characterized in that: It also includes: An iterative calculation module, which is used to take the sequential execution of the wandering behavior module, the summoning behavior module, the besieging behavior module and the individual wolf update behavior module as one round of iteration, set the maximum number of iterations to complete the iterative calculation, output the position of the leading wolf generated in the last iteration, and use this position of the leading wolf as the final multi-UAV task allocation scheme.

3. The multi-UAV task allocation system based on the dynamic wolf pack algorithm according to claim 1, characterized in that: The initialization module is also used to initialize the parameters of the dynamic wolf pack algorithm, specifically including: The population size, the exploring wolf ratio factor, the update ratio factor, and the temperature coefficient in the dynamic wolf pack algorithm.

4. The multi-UAV task allocation system based on dynamic wolf pack algorithm according to claim 1 is characterized by: In the wandering behavior module, the specific method of using the ordered arrangement method to calculate the maximum wandering times of each exploratory wolf in the individual wolf and wandering according to the maximum wandering times is: First, the wolves are sorted in ascending order according to their fitness, and then an integer is randomly generated as the number of wolves to be explored in this iteration according to the formula for the number of wolves to be explored. The formula for the number of wolves to be explored is as follows: Among them, N is the population size, a is the wolf detection ratio factor; Then, according to the task number list and the drone number list, the ordered arrangement method is used to calculate the maximum number of wandering times possnum for each scout wolf. The formula is as follows: in, is the total number of tasks for the mth UAV in the multi-UAV task allocation scheme for the i-th wolf scout; Finally, the fitness of the scout wolf after wandering is calculated. If the fitness is lower than before wandering, the position of the scout wolf is updated. Otherwise, the scout wolf continues to wander until the maximum number of wandering times is reached.

5. The multi-UAV task allocation system based on dynamic wolf pack algorithm according to claim 1 is characterized by: In the summoning behavior module, the specific method of summoning the fierce wolf without changing the task number list is: First, calculate the number of wolves fnum, the calculation formula is as follows: fnum=N–snum–1 Among them, N is the population size, snum is the number of scouting wolves; Then, keep the task number list of the wolf unchanged, randomly select j drone numbers in the drone number list of the wolf as the drone numbers to be modified, and modify the drone numbers to be modified to the drone numbers corresponding to the task numbers corresponding to the drone numbers to be modified in the drone number list of the alpha wolf, 1≤j≤the total number of numbers in the drone number list, and j is a randomly generated integer; Finally, calculate the fitness of the wolf after summoning. If the fitness of the wolf after summoning is lower than that of the alpha wolf, the wolf will replace the alpha wolf as the new alpha wolf and initiate the summoning behavior again according to the new alpha wolf position. If the fitness of the wolf after summoning is lower than that before summoning but higher than that of the alpha wolf, the wolf will update its position. If the fitness of the wolf after summoning is higher than that before summoning, the position of the wolf will remain unchanged.

6. The multi-UAV task allocation system based on dynamic wolf pack algorithm according to claim 1 is characterized by: In the individual wolf update behavior module, the specific method for updating the individual wolf while ensuring that the task number list remains unchanged is: First, an integer is randomly generated as the number L of individual wolf updates in this round of iteration according to the update formula. The update formula is as follows: Among them, N is the population size, b is the update scale factor; Then, under the premise of keeping the task number list unchanged, the drone number list of the L individual wolves ranked in ascending order of fitness values ​​is cleared, and at the same time, in the drone number list of the top X individual wolves in the wolf pack with ascending fitness values, the drone number with the most corresponding tasks is found as the preferred drone number, and the preferred drone number is filled into the drone number list of the L individual wolves ranked in ascending order of fitness values ​​according to the corresponding relationship between the preferred drone number and the task number.

7. The multi-UAV task allocation system based on dynamic wolf pack algorithm according to any one of claims 1 and claims 4-6, characterized in that: The fitness calculation formula is as follows: fitness=α×TCT+β×AFT+γ×ATF Among them, fitness is fitness, α is the weight of the time for the drone group to perform the reconnaissance process, β is the weight of the average flight time of the drone, γ is the weight of the average threat value of the drone, TCT is the time for the drone group to perform the reconnaissance process, AFT is the average flight time of each drone in the drone group, AFT is the average threat value faced by each drone in the drone group during the reconnaissance process, and the values ​​of α and β vary within the constraint condition of α+β=1{α,β∈(0,1)}. When the multi-drone task allocation scheme does not meet the constraint condition, the fitness of the scheme is set to infinity. The reconnaissance process is the process from the starting point to the completion of all tasks and then returning to the starting point; The calculation formula of the time TCT for the drone group to perform the reconnaissance process is as follows: In the formula, represents the time for the mth UAV to perform the reconnaissance process, Nu represents the number of UAVs, The calculation formula is as follows: In the formula, u m represents the mth drone, t k represents the kth task, represents the distance from the mth drone to the location of the first mission t1, represents the distance between the locations of the kth task and the k+1th task, x k represents the location of the kth task, represents the flight distance of the mth UAV to complete the kth mission, Indicates that the mth drone has the last mission t n The distance from the location to the return starting point, represents the flight speed of the mth UAV; The calculation formula for the average flight time AFT of each drone in the drone group is as follows: Where, T m It represents the time required for the mth UAV to complete the reconnaissance process; The calculation formula of the average threat value ATF faced by each drone in the drone group during reconnaissance is as follows: In the formula, It represents the threat coefficient faced by the m-th UAV when performing the k-th mission.

8. The method of multi-UAV task allocation system based on dynamic wolf pack algorithm according to claim 1 is characterized in that: include: The position of each individual wolf in the dynamic wolf pack algorithm is defined as a multi-UAV task allocation scheme consisting of a task number list and a UAV number list, where the task number in the task number list corresponds one-to-one to the UAV number in the UAV number list; The maximum number of wanderings of each exploratory wolf in the individual wolf is calculated by using an ordered arrangement method, and wandering is performed according to the maximum number of wanderings, wherein the wandering is defined as reordering the task numbers in the task number list; The fierce wolf among the individual wolves is made to summon under the premise that the task number list remains unchanged after the task numbers are reordered, and the summoning is defined as the fierce wolf learning the multi-UAV task allocation plan of the current leader wolf; Calculate the corresponding distance probability according to the fitness difference between each individual wolf except the current leader wolf and the current leader wolf among the individual wolves, and carry out siege according to the distance probability, wherein the siege is defined as generating a random number within a preset range, if the distance probability is greater than the random number, the individual wolf mutates, and if the distance probability is less than the random number, the individual wolf learns the multi-UAV task allocation scheme from the individual wolf with a smaller fitness than the individual wolf; After the above-mentioned wandering, summoning and siege, and under the premise of ensuring that the task number list remains unchanged, the positions of individual wolves except the current head wolf are updated to form new positions of all individual wolves, and the current head wolf position among all new positions of individual wolves is output. The position update of the individual wolf is defined as the L individual wolves ranked in ascending order of fitness values ​​learning the multi-UAV task allocation plan of the top X individual wolves ranked in ascending order of fitness values.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by the processor, the multi-UAV task allocation method based on the dynamic wolf pack algorithm described in claim 8 is implemented.

Images