CN115494864A - Heterogeneous multi-unmanned aerial vehicle task allocation method based on ant colony algorithm - Google Patents
Heterogeneous multi-unmanned aerial vehicle task allocation method based on ant colony algorithm Download PDFInfo
- Publication number
- CN115494864A CN115494864A CN202211155444.3A CN202211155444A CN115494864A CN 115494864 A CN115494864 A CN 115494864A CN 202211155444 A CN202211155444 A CN 202211155444A CN 115494864 A CN115494864 A CN 115494864A
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- task
- candidate
- tasks
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 32
- 239000003016 pheromone Substances 0.000 claims abstract description 68
- 239000011159 matrix material Substances 0.000 claims abstract description 61
- 238000012790 confirmation Methods 0.000 claims description 30
- 230000001629 suppression Effects 0.000 claims description 28
- 238000005457 optimization Methods 0.000 claims description 16
- 230000007704 transition Effects 0.000 claims description 16
- 241000257303 Hymenoptera Species 0.000 claims description 15
- 238000003825 pressing Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 230000002238 attenuated effect Effects 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 2
- 238000012545 processing Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 101150039208 KCNK3 gene Proteins 0.000 description 2
- 101150083764 KCNK9 gene Proteins 0.000 description 2
- 238000005056 compaction Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008020 evaporation Effects 0.000 description 1
- 238000001704 evaporation Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000000116 mitigating effect Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000009024 positive feedback mechanism Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
- G05D1/104—Simultaneous control of position or course in three dimensions specially adapted for aircraft involving a plurality of aircrafts, e.g. formation flying
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses a heterogeneous multi-unmanned aerial vehicle task allocation method based on an ant colony algorithm, which classifies current candidate tasks and determines a candidate unmanned aerial vehicle sequence according to classification results and characteristics of different unmanned aerial vehicles; respectively selecting the next task and the undertaker of the next task according to the heuristic information and the pheromone concentration matrix, if all the candidate tasks of the current unmanned aerial vehicle do not meet the constraint condition, reselecting the unmanned aerial vehicle and calculating the selection probability until the selection probability of all the candidate tasks is not all zero; selecting a candidate task to complete one-time distribution, and updating a candidate task sequence and a candidate unmanned aerial vehicle sequence; repeatedly carrying out unmanned aerial vehicle selection, probability calculation and task selection until all tasks are distributed; and updating the pheromone concentration, repeating iteration until the preset iteration times are reached, and determining the optimal distribution scheme according to the optimal ant colony.
Description
Technical Field
The invention relates to the technical field of unmanned aerial vehicle control, in particular to a heterogeneous multi-unmanned aerial vehicle task allocation method based on an ant colony algorithm.
Background
Under the background that modern war has high complexity, unmanned aerial vehicle has become the key development object of many military strong countries by virtue of characteristics such as its tactics realization is nimble, and the ability of continuously fighting is strong, and platform cost is low. In a battlefield background, the types and the scale of tasks are large, a single unmanned aerial vehicle is often difficult to meet requirements, and for the tasks, mutual cooperation and cooperation among heterogeneous unmanned aerial vehicles are needed. Therefore, a good task allocation method is the key for the multi-unmanned aerial vehicle system to exert the maximum utility.
In the related art, deterministic methods are mature. After the mission planning model is established, all feasible solutions in the feasible region can be listed by adopting breadth-first search or depth-first search so as to find the optimal price. In addition, the integer programming method can represent a small-scale task planning problem by establishing an objective function and a constraint condition according to a given purpose and a target. Among them, the hungarian algorithm, the simplex algorithm, the branch-and-bound method, etc. are more commonly used integer programming methods. However, the deterministic method is only suitable for solving a simple model, the complexity of the algorithm is greatly increased along with the increase of the number of unmanned aerial vehicles and tasks, the calculation time is exponentially increased, the efficiency is low, and the fault tolerance is poor.
Compared with a deterministic method, a random method balances the solving precision and the solving time, can obtain an optimal solution or a suboptimal solution in a short time, has certain intelligence and parallelism, has the characteristics of easy realization, low complexity and the like, is suitable for processing large-scale complex problems, and is widely applied in recent years. Especially intelligent algorithms including particle swarm and genetic algorithms. However, as the technology of drones is continuously mature, the classification of drone classes is more and more delicate, and different classes of drones are designed to perform specific tasks. The traditional intelligent algorithm does not consider the isomerism between unmanned aerial vehicles or only considers the isomerism of basic attributes such as speed between the unmanned aerial vehicles, cannot enable the heterogeneous unmanned aerial vehicles to correspond to distributed tasks, and is no longer suitable for modern battlefields.
Disclosure of Invention
In view of the above, the invention provides an ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle task allocation method, which can solve the problem of heterogeneous multi-unmanned aerial vehicle task allocation under the background of complex time sequence requirements among various tasks. And the time sequence requirement is met by using a mode of maintaining the candidate task sequence, and the requirement corresponding to the executed task of the heterogeneous unmanned aerial vehicle is met by using a mode of maintaining the candidate unmanned aerial vehicle sequence.
In order to achieve the purpose, the technical scheme of the invention comprises the following steps:
step 1: initializing ant colonies according to the scale of the task, wherein one ant colony jointly forms a distribution scheme, one ant represents an unmanned aerial vehicle, and any ant in any ant colony has a respective pheromone concentration matrix; meanwhile, initializing parameters of the unmanned aerial vehicle; the drone parameters include the speed of the drone, the drone type, drone coordinates, the type of target, the coordinates of the target, and the type of task.
Step 2: constructing a candidate task sequence aiming at tasks to be executed of the heterogeneous multi-unmanned aerial vehicle system; and constructing a candidate unmanned aerial vehicle sequence according to the types of the current candidate tasks and the characteristics of different types of tasks executed by the heterogeneous unmanned aerial vehicles.
And step 3: and respectively selecting the next task and the undertaker of the next task according to the heuristic information and the pheromone concentration matrix. And if the current unmanned aerial vehicle cannot be matched with all the candidate tasks, deleting the unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence, and repeating the step 3 until the matching is completed.
And 4, step 4: and (4) updating the candidate task sequence, updating the candidate unmanned aerial vehicle sequence, updating the position of the unmanned aerial vehicle, and repeatedly executing the selection of the unmanned aerial vehicles and the distribution of the tasks in the steps (3) and (4) until all the tasks are completely distributed, so as to obtain a distribution matrix of the current iteration.
And 5: and calculating the optimal target value of the current iteration according to the distribution matrix of the current iteration, and updating the optimal distribution matrix by using the distribution matrix of the current iteration if the optimal target value of the current iteration is better than the optimal target value of the optimal distribution matrix.
And updating the pheromone concentration matrix according to the optimized target value of the current iteration, repeatedly executing the step 2 to the step 5 until the set iteration times are reached, obtaining a final optimal distribution matrix, and performing task distribution on the heterogeneous multi-unmanned aerial vehicles by using the final optimal distribution matrix.
Further, a candidate task sequence is constructed for the tasks to be executed of the heterogeneous multi-unmanned aerial vehicle system, and the method specifically comprises the following steps: classifying the tasks to be executed of the current heterogeneous multi-unmanned aerial vehicle system, considering a 3-task scene according to the actual battlefield condition, wherein electromagnetic suppression, attack and confirmation are respectively carried out, and a strict time sequence relationship exists among the 3 tasks, namely: the attack task of any target needs to be carried out within the time range of carrying out electromagnetic suppression on the target, and the confirmation task of any target needs to be carried out after the target is attacked. Therefore, the candidate tasks can be divided into 7 types, namely only the electromagnetic suppression task, only the attack task, only the confirmation task, only the electromagnetic suppression and attack task, only the electromagnetic suppression and confirmation task, only the attack and confirmation task and three tasks. And sequencing the tasks to be executed according to the time sequence relation to obtain a candidate task sequence.
Further, a candidate unmanned aerial vehicle sequence is constructed according to the type of the current candidate task and the characteristics of different types of tasks executed by the heterogeneous unmanned aerial vehicles, and the method specifically comprises the following steps: the heterogeneous unmanned aerial vehicles comprise 4 different types of unmanned aerial vehicles, namely electromagnetic suppression unmanned aerial vehicles, attack unmanned aerial vehicles, reconnaissance unmanned aerial vehicles and helicopters; the electromagnetic suppression unmanned aerial vehicle only executes an electromagnetic suppression task, the attack unmanned aerial vehicle only executes an attack task, the reconnaissance unmanned aerial vehicle only executes a confirmation task, and the helicopter executes the attack task or the confirmation task. And constructing a reasonable candidate unmanned aerial vehicle sequence to enable the sequence to correspond to the candidate tasks.
Furthermore, the size of the pheromone concentration matrix is NxN +1, N is the number of target points, the Nth column represents the pheromone concentration of the unmanned aerial vehicle heading to each target from the Nth target point, and the (N + 1) th column represents the pheromone concentration of the unmanned aerial vehicle heading to each target from the initial position; the heuristic information is the distance between the current position of the unmanned aerial vehicle and the target position.
Further, in step 3, the next task and the undertaker of the next task are respectively selected according to the heuristic information and the pheromone concentration matrix. When the last task is not an electromagnetic suppression task of any target, randomly selecting one unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence as a bearer of the next task, acquiring the position of the current unmanned aerial vehicle, traversing all candidate tasks, calculating the transition probability of the candidate tasks according to the heuristic information and the pheromone concentration matrix, and selecting the next task for all candidate tasks matched with the current unmanned aerial vehicle in a roulette mode. When the last task is an electromagnetic suppressing task for any target, the task to be executed at the moment is designated as an attack task for the target instead of selecting from the candidate task sequence in a roulette mode. Traversing all the candidate unmanned aerial vehicles, calculating the transition probability of the candidate unmanned aerial vehicles according to the heuristic information and the pheromone concentration matrix, and selecting the undertaker of the next task by adopting a roulette mode for all the candidate unmanned aerial vehicles matched with the current task; after the attack task is completed, the candidate task sequence is updated as usual.
Further, if the current unmanned aerial vehicle and all candidate tasks cannot be matched, deleting the unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence, specifically: the same unmanned aerial vehicle can only execute one task at the same target point; recording the target point reached by each unmanned aerial vehicle, setting the transition probability of selecting the candidate task as 0 if the current unmanned aerial vehicle executes other tasks at the target point corresponding to the candidate task in the process of traversing the candidate task, deleting the unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence when the transition probabilities of the unmanned aerial vehicle and all the candidate tasks are 0, namely that the current unmanned aerial vehicle and all the candidate tasks cannot be matched, and selecting a new unmanned aerial vehicle as a bearer of the next task from the updated candidate unmanned aerial vehicle sequence.
Further, updating the pheromone concentration matrix according to the optimized target value of the current iteration, specifically: during updating, firstly, pheromones are volatilized, namely, the pheromone concentrations of all nodes are attenuated in a certain proportion, and then, the pheromone concentrations of the nodes which are walked by ants, namely different target points which are reached by the unmanned aerial vehicle, are strengthened in a certain proportion.
Further, the optimization target value is the total time of the task or the profit value; if the optimization target value is the total task time, the total task time is preferably small: if the optimization target value is the profit value, the profit value is greatly excellent.
Has the advantages that:
1. the invention provides an ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle task allocation method which is used for solving the problem of heterogeneous multi-unmanned aerial vehicle task allocation under the background that complex time sequence requirements exist among various tasks. And the time sequence requirement is met by using a mode of maintaining the candidate task sequence, and the requirement corresponding to the executed task of the heterogeneous unmanned aerial vehicle is met by using a mode of maintaining the candidate unmanned aerial vehicle sequence.
2. The heterogeneous multi-unmanned aerial vehicle task allocation method based on the ant colony algorithm has strong generalization and can be flexibly applied to various scenes. The heuristic information is fully utilized and a positive feedback mechanism enables the heuristic information to quickly approach the optimal solution.
Detailed Description
The present invention will be described in detail below with reference to examples.
The invention provides an ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle task allocation method, which comprises the following steps:
the ant colony is initialized according to the task scale, one ant colony jointly forms a distribution scheme, one ant represents one unmanned aerial vehicle, in order to improve the search efficiency and reduce the mutual interference among the unmanned aerial vehicles, any ant in any ant colony has a respective pheromone concentration matrix, and the size of the matrix is determined by the target number. And simultaneously initializing parameters such as the speed of the unmanned aerial vehicle, the type of the unmanned aerial vehicle, the coordinates of the unmanned aerial vehicle, the type of the target, the coordinates of the target, the type of the task and the like.
And constructing a candidate unmanned aerial vehicle matrix according to the type of the current candidate task and the characteristics of different types of tasks executed by the heterogeneous unmanned aerial vehicles.
Randomly selecting one unmanned aerial vehicle from the candidate unmanned aerial vehicle matrix as a bearer of the next task, acquiring the position of the current unmanned aerial vehicle, traversing all candidate tasks, calculating the selection probability of each candidate task according to the heuristic information and the pheromone concentration, and selecting the next task by adopting a roulette mode. And if the current unmanned aerial vehicle cannot be matched with all the candidate tasks, deleting the unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence, and repeating the step.
And updating the candidate task matrix, updating the position of the unmanned aerial vehicle, and judging whether all tasks are completed.
If not, selecting the next task by adopting a roulette mode repeatedly according to the heuristic information and the pheromone concentration.
And updating the pheromone concentration matrix according to the total time required for completing all tasks, and updating the optimal distribution matrix.
In one embodiment of the invention, a 3-mission scenario is considered based on actual battlefield conditions, with the 3 missions being electromagnetic suppression, attack, and confirmation, respectively.
The tasks have strict time sequence relation, the attack task for any target needs to be carried out within the time range of carrying out electromagnetic suppression on the target, and the confirmation task for any target needs to be carried out after the attack task for the target is completed.
The candidate tasks are divided into 7 types according to the composition of the tasks in the current candidate tasks, wherein the tasks are only electromagnetic suppression tasks, only attack tasks, only confirmation tasks, only electromagnetic suppression and attack tasks, only electromagnetic suppression and confirmation tasks, only attack and confirmation tasks and three tasks.
In one embodiment of the invention, a sequence of candidate drones is constructed according to the type of candidate task. There are 3 different types of unmanned aerial vehicles, be electromagnetism suppression unmanned aerial vehicle respectively, attack unmanned aerial vehicle, reconnaissance unmanned aerial vehicle and helicopter.
The electromagnetic suppression unmanned aerial vehicle can execute an electromagnetic suppression task as far as possible, the attack unmanned aerial vehicle can only execute the attack task, and the reconnaissance unmanned aerial vehicle can only execute a confirmation task. The helicopter can perform both attack and confirmation tasks.
A reasonable sequence of candidate drones needs to be constructed to correspond to the candidate tasks.
For example, when the candidate task only comprises the electromagnetic suppression task, all the electromagnetic suppression unmanned aerial vehicles form a candidate unmanned aerial vehicle sequence; when only the attack and confirmation tasks exist in the candidate tasks, all attack unmanned planes with attack or confirmation capabilities, the reconnaissance unmanned plane and the helicopter form a candidate unmanned plane sequence.
In one embodiment of the invention, after the sequence of candidate drones is determined, a bearer is randomly selected from the sequence as the bearer for the next task.
After the unmanned aerial vehicle is selected, traversing the candidate task, and calculating the transition probability according to heuristic information and the pheromone concentration matrix, wherein the heuristic information is the distance between the current position and the target position of the unmanned aerial vehicle. In the traversal process, if the candidate task and the heterogeneous unmanned aerial vehicle cannot be matched, if the attack task is allocated to the electromagnetic suppression unmanned aerial vehicle, the state transition probability of the task is set to be 0.
When the same unmanned aerial vehicle can only execute one task at the same target point, and when the candidate task is traversed, if the current unmanned aerial vehicle executes other tasks at the target point corresponding to the candidate task, the state transition probability of the task is also set to be 0.
And after traversing all the candidate tasks, selecting the next task by adopting a roulette mode according to the state transition probability of each selected task.
And updating the candidate tasks, for example, after the attack task on a certain target point is completed, deleting the task from the candidate tasks, and adding the confirmation task on the target point into the candidate task sequence.
When the last task is not an electromagnetic suppression task of any target, randomly selecting one unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence as a bearer of the next task, acquiring the position of the current unmanned aerial vehicle, traversing all candidate tasks, calculating the transition probability of the candidate tasks according to the heuristic information and the pheromone concentration matrix, and selecting the next task for all candidate tasks matched with the current unmanned aerial vehicle in a roulette mode.
In one embodiment of the present invention, since the electromagnetic compaction task on a target needs to be continued until the attack on the target is completed, in order to achieve maximum utilization of time and resources, the electromagnetic compaction task and the attack task on the same target need to be scheduled as soon as possible at the same time. Therefore, when selecting the task, if the previous task is the electromagnetic suppressing task for a certain target, the task to be executed at the moment is designated as the attacking task for the target instead of selecting from the candidate task sequence in the manner of roulette. Traversing all the candidate unmanned aerial vehicles, calculating the transition probability of the candidate unmanned aerial vehicles according to the heuristic information and the pheromone concentration matrix, and selecting the undertaker of the next task by adopting a roulette mode for all the candidate unmanned aerial vehicles matched with the current task; after the attack task is completed, the candidate task sequence is updated as usual.
In one embodiment of the invention, the selection of the task and the updating of the candidate task are repeated until all tasks are executed, that is, a feasible solution is obtained. The feasible solution includes the order of executing the tasks and the corresponding relationship between the unmanned aerial vehicle and the tasks, so that the total time under the allocation scheme can be obtained.
In one embodiment of the present invention, the pheromone concentration may be updated after a feasible solution is found.
Firstly, the pheromone is volatilized, and the concentration of all the pheromones is attenuated according to a certain proportion.
And then, the concentration of the pheromone is strengthened, the ants leave the pheromone in the passing path, and the shorter the path is, the stronger the concentration of the left pheromone is. In the example, the target point and the target point, and the target point and the initial point of the unmanned aerial vehicle are communicated in pairs to form a graph, the unmanned aerial vehicle selects one edge in the graph according to the assigned task and the current position, and the pheromone concentration of the selected edge is enhanced.
When the optimization goal is that the total time is shortest, the pheromone concentration of the selected side is higher under the scheme with shorter total time, and the probability that the scheme with shorter total time is selected is higher under the action of positive feedback.
In one embodiment of the invention, in addition to minimizing task time, revenue maximization may also be used as an optimization goal for task allocation. Different types of targets have different strategic values in view of actual combat situations, for example, the strategic value of a command center is higher than that of a common radar. In addition, targets of the same type also have different strategic values, for example, a radar located at a transportation junction has a higher strategic value than other radars.
After electromagnetic suppression, attack and confirmation tasks are completed on a target, the target can be considered to lose the strategic value completely, the product of the strategic value and the effective working time of the target can be considered as the total profit provided by the target for the enemy, and the smaller the total profit provided by the opposite enemy is, the larger the total profit is. Therefore, during the battle, the higher profit can be obtained by preferentially aiming at the target with high strategic value.
When the optimization goal is that the total time is shortest, the intensity of pheromone concentration enhancement is inversely proportional to the total time of the scheme when the pheromone concentration is updated. Similarly, when the optimization target is the total profit, and the pheromone concentration is updated, the total time of the original scheme is only replaced by the total profit of the opposite enemy, and the probability that the scheme with the highest total profit is selected is higher under the action of positive feedback.
The task allocation problem is modeled as a finite buffer flow shop scheduling problem. The model is often used for plant pipeline optimization problems and can be described as N workpieces being machined on M machines, each workpiece requiring K times to be machined by the machine. Under the problem of task allocation of the unmanned aerial vehicle, the unmanned aerial vehicle is represented as a machine, the target to be completed is a workpiece, the task to be completed is the processing frequency, and the specific processing task can only be completed by the specific machine.
Based on the problem that the M unmanned aerial vehicles execute K tasks to N targets, the following assumptions are provided:
1) Each target requires the drone to perform multiple tasks such as electromagnetic mitigation, identification, attack, and validation. Under certain simplified conditions, the three tasks of electromagnetic suppression, attack and confirmation can be changed;
2) Different drones have the ability to perform different tasks (wherein the helicopter can perform both the attack and the confirmation tasks, the electromagnetically-suppressed drone only performs the electromagnetically-suppressed task, the attacking drone only can perform the attack task, and the reconnaissance drone only can perform the confirmation task.
3) The number of unmanned aerial vehicles is greater than or equal to the target number.
Therefore, on the premise of meeting the relevant constraint conditions, the objective function of the problem is to minimize the maximum task completion time J. In combination with the model of the flow shop scheduling problem, the constraint optimization model is defined as follows
The specific meanings of the variables are shown in table 1:
TABLE 1 meanings of variables
Equation (1) is an optimization objective function that minimizes the maximum task completion time when constraint (2) is satisfied. J (| C) N,M The value of (-) is the cumulative time matrix C N,M Of (2) is calculated. The specific definitions of the variables and the set of constraints Γ please see below.
Time-of-flight cost matrix T ij
T ij The method is a matrix with N rows and N + M columns, wherein the rows represent targets in the task allocation problem, the front N columns are targets, and the rear M columns are unmanned planes. Element t ij Representing the time of flight from node i to node j. The specific definition is as follows:
task accumulation time matrix C N,M
C N,M And accumulating a matrix of time spent on completing the task for the unmanned aerial vehicle. The rows of the matrix represent objects and the columns represent noneAnd (4) man-machine. Element c n,m The current time when the mth unmanned aerial vehicle completes a certain task at the nth target.
Wherein a current time c when a task of the nth target is completed by the mth drone n,m Given by:
C n,m ＝max{C n ,C m } (4)
C n ＝max{C n,i |i＝1,2,…,M} (5)
C m ＝t i,j +max{C j,m |j＝1,2,…,N},t i,j ∈T ij (6)
C n is C N,M The maximum value in the nth row of (a) represents the maximum time required for the unmanned aerial vehicle to wait for the completion of other tasks on the target before completing the current task of the target n; c m Is composed of two parts, the former part is t i,j Time required for the mth drone to go to the nth target to execute a new task, the latter part being C N,M The maximum value in the mth column of (a) represents that the unmanned aerial vehicle can go to the target n to execute a new task after all tasks before the unmanned aerial vehicle processes the unmanned aerial vehicle.
For matrix C N,M The maximum of all its elements represents the time eventually needed to complete the entire assignment scheme.
Target value matrix V N
In addition to minimizing the maximum task performance, the method also introduces the index of maximizing the operational profit index. The index designs a target value matrix V by considering different strategic values of different target points N Higher gains can be achieved by striking targets of high strategic value first.
Equation (7) is the target value matrix, V i The strategic value of target i; c n Defining equation (5), accumulating time matrix C for the task n,m The largest value in the nth row represents the target effective operating time. The sum of products of the strategic value and the effective operating time of the targets is minimized by a reasonable distribution. The overall optimization objective formula is expressed as follows:
target task sequence O NK
O NK The order in which all targets in the assignment scheme are executed. Each element of the sequence is the number of the object and the task number is determined by the number of times the object appears. Such as the element o i-k Represents O NK The kth occurrence of the ith target represents the kth task. Taking a case where there are only two targets and two tasks as an example, as shown in Table 2, the 3 rd element O NK (3) Is 1 and is at O NK In the 2 nd occurrence, so the corresponding operation is o 1-2 It means that the 2 nd task of the 1 st object is to be executed.
TABLE 2 task sequence definitions
| O NK | O NK (1) | O NK (2) | O NK (3) | O NK (4) |
| Target | 1 | 2 | 1 | 2 |
| Task | 0 1-1 | 0 2-1 | 0 1-2 | 0 2-2 |
Unmanned aerial vehicle distribution matrix pi K,N
Π K,N A matrix is assigned to the drones. The rows of the matrix correspond to tasks that the target needs to execute, and the columns correspond to the targets. Element pi k,n The k-th task representing the nth target is formed by an unmanned plane pi k,n And (6) executing. II type K,N Definition of (c):
constraint definition
According to the unmanned aerial vehicle assignment model defined above, the constraints can be divided according to strong constraints and weak constraints, and the specific division is as shown in the following table:
TABLE 3 constraint definitions
Constraint processing
In the above classification of the constraints, the weak constraint refers to a constraint condition that has been taken into consideration in the modeling process and is satisfied in the model. For example, for constraint Γ 1 Unmanned aerial vehicle distribution matrix pi K,N Satisfies the requirement that all tasks are executedSolving; for constraint Γ 2 ，∏ K,N The relationship between the drone and the target task is limited.
For strong constraints, the process proceeds by selecting a method to check and adjust the potential solutions in each iteration cycle and make them meet all constraints.
Pheromone density matrix P
The ant colony algorithm takes ants as main bodies, takes pheromones as information transmitted between iterations, and abstracts paths among foods into a graph. At the beginning, the pheromone concentration on all paths is the same, after a period of time, ants on shorter paths have high round-trip speed, the ants participating in the path also gradually change, more pheromones are left, and the probability of the ants selecting the path is higher; the longer path has fewer ants and slow back-and-forth speed, and the pheromone concentration is less and less in continuous volatilization. Under the action of positive feedback, the ant colony gradually converges to the optimal path. In the task allocation problem, one ant colony represents one unmanned aerial vehicle cluster, and one ant is one unmanned aerial vehicle. Between the target, unmanned aerial vehicle and target two by two communicate and constitute an undirected graph, and the pheromone concentration of every limit on the graph constitutes pheromone concentration matrix P.
To prevent interference between ants in a same ant colony, each ant has its own pheromone concentration matrix P k ,
And when i = N +1, the representative unmanned aerial vehicle is located at the initial point and transfers to other target points.
Ant colony coding design
For the assignment problem that M unmanned aerial vehicles go to N targets to execute K tasks, each ant colony group X is designed i Is composed of M pieces of antEach ant comprises X it Representing a UAV, element x in the vector it And (n) is an integer, and the target to which the unmanned aerial vehicle t needs to go and the task correspondingly executed by the target can be obtained after conversion, and the target is referred to as a target city for short. As defined in formulae (10) and (11):
the constraint of equation (11) indicates that each group of ants accomplishes all the tasks for all the goals. Equation (10) can be transformed to pi for simplicity of programming, for ease of comparison with other algorithms K,N The flow chart is shown in table 4:
table 4 pi K,N Transformation algorithm
The specific process can refer to the example given below, which describes X when 3 target 4-rack drones perform 2 tasks i Corresponding pi K,N ：
Ant group X i Examples of the invention
| X i | x i (1) | x i (2) |
| X i1 | 4 | 5 |
| X i2 | 1 | |
| X i3 | 2 | |
| X i4 | 3 | 6 |
Ant group X i Corresponding pi K,N
For assignment time sequence O NK The acquisition mode adopts a sequence selection strategy following ant activities, and after the UAV and a task-city to be executed by the UAV are selected, O is recorded at the same time NK In (1).
| O NK | O NK (1) | O NK (2) | O NK (3) | O NK (4) | O NK (5) | O NK (6) |
| Target | 1 | 1 | 2 | 2 | 2 | 1 |
| Task | 0 1-1 | 0 2-1 | 0 1-2 | 0 2-2 | 0 2-3 | 0 3-3 |
Assigning ant colony algorithm designs
Classifying the current candidate tasks, constructing a candidate unmanned aerial vehicle sequence according to different classification results and the property that the heterogeneous unmanned aerial vehicle can only execute specific tasks, and randomly selecting one candidate unmanned aerial vehicle sequence as a bearer of the next task. And traversing all the candidate tasks to respectively calculate the transition probability, and selecting the next task by adopting a roulette mode. For example, at the very beginning of the assignment, the set of electromagnetic squashing tasks of all targets constitutes a candidate task sequence, whereas all electromagnetic squashing drones constitute a candidate drone sequence, since the electromagnetic squashing tasks must first be executed. And only confirmation tasks for certain tasks are left at the end of the allocation, and all reconnaissance drones and helicopters with the capability form a candidate airplane sequence.
When the ants select the targets going to next step, the transfer probability is calculated according to the pheromone concentration and the heuristic information, and the following formula is specifically adopted:
wherein: i is the current target of the ant, and j is the target to be accessed; the able in the formula (12) represents that the unmanned aerial vehicle can execute the type of task, and other tasks are not executed at the target before, so that the heterogeneous unmanned aerial vehicle is ensured not to select tasks which cannot be completed by the heterogeneous unmanned aerial vehicle, and meanwhile, the constraint that the same helicopter can only execute one task at the same target is met. Eta in formula (13) ij To enlighten the message, it takes t to go to the next node j i,j The reciprocal of (a), i.e. ants tend to select nodes close to themselves for state transition; c in the formula (14) is an initialization pheromone concentration and is an adaptive value, tau ij (t) pheromone strength of nodes from i to j at t moment; allowed k Is a candidate task sequence; α, β are weighted values of pheromones and visibility.
After the assignment is completed once, the candidate task is updated, for example, after the attack task on the target 1 is completed, the task is deleted from the candidate task sequence, the confirmation task on the target is added, the candidate unmanned aerial vehicle sequence is also updated along with the update of the candidate task, and meanwhile, the accumulated time matrix is updated. And repeating the assignment and the updating until all the tasks are executed, thus obtaining an assignment scheme.
Updating pheromone concentration
Firstly, the pheromone is volatilized, namely the pheromone concentration of all the nodes is attenuated in a certain proportion, then the pheromone concentration of the nodes which are walked by the ants, namely different target points which are reached by the unmanned aerial vehicle is strengthened, the shorter the required total time is, the more pheromones are left by the ants, and the stronger the pheromone concentration is. In order to improve the search efficiency and reduce the interference between drones, a separate pheromone matrix SP is designed for each ant in the ant group, i.e. each drone M . The pheromone is also updated with the iteration, and the update formula is as follows:
τ ij ∈SP M (15)
wherein: m is the number of ants, rho is the element (0, 1), and is the pheromone evaporation rate, delta tau ij k Pheromones left for the kth ant group on paths i to j, C k The total path of the kth ant group after the path is completed is the total time required under the allocation scheme in this example.
For the drone assignment problem, the specific flow of the whole algorithm is shown in table 5:
TABLE 5 Ant colony assignment algorithm for unmanned aerial vehicle
In the following, the ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle task allocation method is further explained through a simulation experiment.
The task area is set to be N =4 in number of targets and M =5 in number of total airplanes, wherein 2 unmanned aerial vehicles are electromagnetically pressed, 1 unmanned aerial vehicle is attacked, 1 unmanned aerial vehicle is reconnaissance, 1 helicopter is helicopter, and K =3 in number of tasks are respectively an electromagnetic pressing task, an attacking task and a confirming task. The electromagnetic pressing unmanned aerial vehicle can only execute an electromagnetic pressing task, the attacking unmanned aerial vehicle can only execute an attacking task, the reconnaissance unmanned aerial vehicle can only execute a confirmation task, the helicopter can execute the attacking task and the confirming task, the attacking task can be completed only when the target is electromagnetically pressed, and the confirmation task is executed after the attacking task.
TABLE 6
| UAV | SP | Type | V o |
| 1 | (9,10,50) | H | 1.0 |
| 2 | (18,10,50) | E | 1.5 |
| 3 | (27,10,50) | E | 1.5 |
| 4 | (36,10,50) | A | 1.5 |
| 5 | (45,10,50) | S | 1.5 |
TABLE 7
| Target | EP | V a |
| 1 | (10,1.4,140) | 1 |
| 2 | (20,3.6,140) | 2 |
| 3 | (30,7.7,140) | 3 |
| 4 | (40,11.4,140) | 4 |
In this example, the specific parameters are shown in table 6, where the starting position of the drone/helicopter is SP and the Type is Type, where H represents the helicopter, E represents the electromagnetically-suppressed drone, a represents the attacking drone, S represents the reconnaissance drone, and the flying speed is V o Continuing with Table 2, the coordinate position of the target is EP and the value of the target is V a . In the simulation, the ant colony number is set as, the iteration number is set as, and the weight coefficient is set as
Simulation experiment I:
setting the optimization target to be the shortest total time, performing task allocation through the algorithm, and enabling allocation results to be shown in the following
| Π K,N | Target1 | Target2 | Target3 | Target4 |
| Task1 | 3 | 3 | 3 | 2 |
| Task2 | 4 | 4 | 4 | 4 |
| Task3 | 5 | 1 | 5 | 5 |
| O NK | O NK (1) | O NK (2) | O NK (3) | O NK (4) | O NK (5) | O NK (6) |
| Target | 4 | 4 | 4 | 3 | 3 | 2 |
| Task | 0 4-1 | 0 4-2 | 0 4-3 | 0 3-1 | 0 3-2 | 0 2-1 |
| O NK | O NK (7) | O NK (8) | O NK (9) | O NK (10) | O NK (11) | O NK (12) |
| Target | 2 | 1 | 1 | 3 | 2 | 1 |
| Task | 0 2-2 | 0 1-1 | 0 1-2 | 0 3-3 | 0 2-3 | 0 1-3 |
It can be seen that as ant populations are continuously updated, the total time required is less and less, and the distribution results satisfy various constraints. The effectiveness of the algorithm was demonstrated.
And (2) simulation experiment II:
setting the optimization target to be the highest total income, carrying out task allocation through the algorithm, and enabling allocation results to be shown in the specification
| Π K,N | Target1 | Target2 | Target3 | Target4 |
| Task1 | 2 | 3 | 3 | 3 |
| Task2 | 4 | 4 | 4 | 4 |
| Task3 | 5 | 1 | 5 | 5 |
| O NK | O NK (1) | O NK (2) | O NK (3) | O NK (4) | O NK (5) | O NK (6) |
| Target | 4 | 4 | 4 | 3 | 3 | 3 |
| Task | 0 4-1 | 0 4-2 | 0 4-3 | 0 3-1 | 0 3-2 | 0 3-3 |
| O NK | O NK (7) | O NK (8) | O NK (9) | O NK (10) | O NK (11) | O NK (12) |
| Target | 2 | 2 | 1 | 1 | 2 | 1 |
| Task | 0 2-1 | 0 2-2 | 0 1-2 | 0 2-2 | 0 2-3 | 0 2-3 |
In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. The heterogeneous multi-unmanned aerial vehicle task allocation method based on the ant colony algorithm is characterized by comprising the following steps of:
step 1: initializing ant colonies according to the scale of the task, wherein one ant colony jointly forms a distribution scheme, one ant represents an unmanned aerial vehicle, and any ant in any ant colony has a respective pheromone concentration matrix; initializing parameters of the unmanned aerial vehicle; the unmanned aerial vehicle parameters comprise the speed of the unmanned aerial vehicle, the type of the unmanned aerial vehicle, the coordinates of the unmanned aerial vehicle, the type of a target, the coordinates of the target and the type of a task;
step 2: aiming at tasks to be executed by a heterogeneous multi-unmanned aerial vehicle system, constructing a candidate task sequence; constructing a candidate unmanned aerial vehicle sequence according to the types of the current candidate tasks and the characteristics of different types of tasks executed by the heterogeneous unmanned aerial vehicles;
and step 3: respectively selecting a next task and a bearer of the next task according to the heuristic information and the pheromone concentration matrix;
if the current unmanned aerial vehicle and all the candidate tasks cannot be matched, deleting the unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence, and repeating the step 3 until the matching is completed;
and 4, step 4: updating the candidate task sequence, updating the candidate unmanned aerial vehicle sequence, updating the position of the unmanned aerial vehicle, repeatedly executing the selection of the unmanned aerial vehicle in the steps 3 and 4 and the distribution of the tasks until all the tasks are distributed, and obtaining a distribution matrix of the current iteration;
and 5: calculating an optimal target value of the current iteration according to the allocation matrix of the current iteration, and updating the optimal allocation matrix by using the allocation matrix of the current iteration if the optimal target value of the current iteration is better than the optimal target value of the optimal allocation matrix;
and updating the pheromone concentration matrix according to the optimized target value of the current iteration, repeatedly executing the step 2 to the step 5 until the set iteration times are reached, obtaining a final optimal distribution matrix, and performing task distribution on the heterogeneous multi-unmanned aerial vehicles by using the final optimal distribution matrix.
2. The ant colony algorithm-based heterogeneous multi-drone multitask allocation method according to claim 1, wherein the candidate task sequence is constructed for the tasks to be executed by the heterogeneous multi-drone system, specifically comprising the steps of:
the method comprises the steps of classifying tasks to be executed of the current heterogeneous multi-unmanned aerial vehicle system, considering a 3-task scene according to actual battlefield conditions, wherein electromagnetic suppression, attack and confirmation are respectively adopted, and a strict time sequence relationship exists among the 3 tasks, namely: the attack task of any target needs to be carried out within the time range of carrying out electromagnetic suppression on the target, and the confirmation task of any target needs to be carried out after the target is attacked; therefore, the candidate tasks can be divided into 7 types, namely, only electromagnetic pressing tasks, only attack tasks, only confirmation tasks, only electromagnetic pressing and attacking tasks, only electromagnetic pressing and confirming tasks, only attack and confirming tasks and three tasks;
and sequencing the tasks to be executed according to the time sequence relation to obtain a candidate task sequence.
3. The ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle multitask allocation method according to claim 2, wherein the candidate unmanned aerial vehicle sequence is constructed according to the type of the current candidate task and the characteristics of different types of tasks executed by heterogeneous unmanned aerial vehicles, and specifically comprises the following steps:
the heterogeneous unmanned aerial vehicles comprise 4 different types of unmanned aerial vehicles, namely electromagnetic pressing unmanned aerial vehicles, attack unmanned aerial vehicles, reconnaissance unmanned aerial vehicles and helicopters; the electromagnetic suppression unmanned aerial vehicle only executes an electromagnetic suppression task, the attack unmanned aerial vehicle only executes an attack task, the reconnaissance unmanned aerial vehicle only executes a confirmation task, and the helicopter executes the attack task or the confirmation task;
and constructing a reasonable candidate unmanned aerial vehicle sequence to enable the sequence to correspond to the candidate tasks.
4. The ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle multitask allocation method according to claim 3, characterized in that the pheromone concentration matrix size is N x N +1, N is the number of target points, the nth column represents the pheromone concentration of the unmanned aerial vehicle heading for each target at the nth target point, and the N +1 column represents the pheromone concentration of the unmanned aerial vehicle heading for each target at the initial position;
and the heuristic information is the distance between the current position of the unmanned aerial vehicle and the target position.
5. The ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle multitask allocation method according to claims 1-4, wherein in the step 3, the next task and the undertaker of the next task are respectively selected according to heuristic information and an pheromone concentration matrix;
when the previous task is not an electromagnetic pressing task of any target, randomly selecting an unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence as a bearer of the next task, acquiring the position of the current unmanned aerial vehicle, traversing all candidate tasks, calculating the transition probability of the candidate tasks according to heuristic information and an pheromone concentration matrix, and selecting the next task for all candidate tasks matched with the current unmanned aerial vehicle in a roulette mode;
when the last task is an electromagnetic suppression task for any target, selecting from a candidate task sequence in a roulette mode, and designating the task to be executed at the moment as an attack task for the target; traversing all the candidate unmanned aerial vehicles, calculating the transition probability of the candidate unmanned aerial vehicles according to the heuristic information and the pheromone concentration matrix, and selecting the undertaker of the next task by adopting a roulette mode for all the candidate unmanned aerial vehicles matched with the current task; after the attack task is completed, the candidate task sequence is updated as usual.
6. The ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle multitask allocation method according to claim 5, wherein if the current unmanned aerial vehicle and all candidate tasks cannot be matched, the unmanned aerial vehicle is deleted from the candidate unmanned aerial vehicle sequence, specifically:
the same unmanned aerial vehicle can only execute one task at the same target point; recording the target point reached by each unmanned aerial vehicle, setting the transition probability of selecting the candidate task as 0 if the current unmanned aerial vehicle executes other tasks at the target point corresponding to the candidate task in the process of traversing the candidate task, deleting the unmanned aerial vehicle from the candidate unmanned aerial vehicle sequence when the transition probabilities of the unmanned aerial vehicle and all the candidate tasks are 0, namely that the current unmanned aerial vehicle and all the candidate tasks cannot be matched, and selecting a new unmanned aerial vehicle as a bearer of the next task from the updated candidate unmanned aerial vehicle sequence.
7. The ant colony algorithm-based heterogeneous multi-drone multitask allocation method according to any one of claims 1, 2, 3, 4 or 6, wherein the pheromone concentration matrix is updated according to the optimal target value of the current iteration, specifically:
during updating, firstly, pheromones are volatilized, namely, the pheromone concentrations of all nodes are attenuated in a certain proportion, and then, the pheromone concentrations of the nodes which are walked by ants, namely different target points which are reached by the unmanned aerial vehicle, are strengthened in a certain proportion.
8. The ant colony algorithm-based heterogeneous multi-unmanned aerial vehicle multitask allocation method according to claim 1, characterized in that the optimization target value is a total mission time or a profit value;
if the optimization target value is the total task time, the total task time is preferably small:
and if the optimization target value is the profit value, the profit value is better.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211155444.3A CN115494864A (en) | 2022-09-22 | 2022-09-22 | Heterogeneous multi-unmanned aerial vehicle task allocation method based on ant colony algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211155444.3A CN115494864A (en) | 2022-09-22 | 2022-09-22 | Heterogeneous multi-unmanned aerial vehicle task allocation method based on ant colony algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115494864A true CN115494864A (en) | 2022-12-20 |
Family
ID=84469823
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211155444.3A Pending CN115494864A (en) | 2022-09-22 | 2022-09-22 | Heterogeneous multi-unmanned aerial vehicle task allocation method based on ant colony algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115494864A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115689252A (en) * | 2022-12-28 | 2023-02-03 | 杭州中安电子有限公司 | Task allocation method and system based on self-adaptive ant colony algorithm |
| CN118092361A (en) * | 2024-04-17 | 2024-05-28 | 哈尔滨工业大学（威海） | Heterogeneous multi-robot task scheduling method and system for textile industry |
-
2022
- 2022-09-22 CN CN202211155444.3A patent/CN115494864A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115689252A (en) * | 2022-12-28 | 2023-02-03 | 杭州中安电子有限公司 | Task allocation method and system based on self-adaptive ant colony algorithm |
| CN118092361A (en) * | 2024-04-17 | 2024-05-28 | 哈尔滨工业大学（威海） | Heterogeneous multi-robot task scheduling method and system for textile industry |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115494864A (en) | Heterogeneous multi-unmanned aerial vehicle task allocation method based on ant colony algorithm | |
| CN111240353B (en) | Unmanned aerial vehicle collaborative air combat decision method based on genetic fuzzy tree | |
| CN111859541B (en) | PMADDPG multi-unmanned aerial vehicle task decision method based on transfer learning improvement | |
| CN114281104B (en) | Multi-unmanned aerial vehicle cooperative regulation and control method based on improved ant colony algorithm | |
| Ming et al. | Improved discrete mapping differential evolution for multi-unmanned aerial vehicles cooperative multi-targets assignment under unified model | |
| CN112633654A (en) | Multi-unmanned aerial vehicle task allocation method based on improved cluster expansion consistency bundle algorithm | |
| CN114020031B (en) | Unmanned aerial vehicle cluster collaborative dynamic target searching method based on improved pigeon colony optimization | |
| CN115755963B (en) | Unmanned aerial vehicle group collaborative mission planning method considering carrier delivery mode | |
| CN114594794A (en) | Multi-machine collaborative task planning method considering subsystem execution capacity | |
| CN115047907B (en) | Air isomorphic formation command method based on multi-agent PPO algorithm | |
| Fu et al. | A Two‐Layer Task Assignment Algorithm for UAV Swarm Based on Feature Weight Clustering | |
| CN115494873A (en) | Heterogeneous multi-unmanned aerial vehicle cooperative task allocation method based on Monte Carlo tree search architecture under time sequence constraint | |
| CN111797966A (en) | Multi-machine cooperative global target distribution method based on improved sheep swarm algorithm | |
| Wu et al. | Heterogeneous mission planning for multiple uav formations via metaheuristic algorithms | |
| CN113741186B (en) | Double-aircraft air combat decision-making method based on near-end strategy optimization | |
| CN113608546B (en) | Unmanned aerial vehicle group task distribution method based on quantum sea lion mechanism | |
| Jing et al. | Cooperative task assignment for heterogeneous multi-UAVs based on differential evolution algorithm | |
| CN116088586B (en) | Method for planning on-line tasks in unmanned aerial vehicle combat process | |
| Gaowei et al. | Using multi-layer coding genetic algorithm to solve time-critical task assignment of heterogeneous UAV teaming | |
| CN116595864A (en) | Multi-unmanned aerial vehicle task allocation method based on mean shift clustering algorithm optimization | |
| He et al. | An operation planning generation and optimization method for the new intelligent combat SoS | |
| Yeding et al. | Research on UAV task assignment based on hybrid particle swarm algorithm | |
| Shuo et al. | Research on distributed task allocation of loitering munition swarm | |
| Shi et al. | The UAV swarm mission planning based on adaptive genetic algorithm | |
| Wei et al. | UCAVs online collaborative path planning method based on dynamic task allocation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |