CN117590757B - Multi-unmanned aerial vehicle cooperative task allocation method based on Gaussian distribution sea-gull optimization algorithm - Google Patents
Multi-unmanned aerial vehicle cooperative task allocation method based on Gaussian distribution sea-gull optimization algorithm Download PDFInfo
- Publication number
- CN117590757B CN117590757B CN202410079406.7A CN202410079406A CN117590757B CN 117590757 B CN117590757 B CN 117590757B CN 202410079406 A CN202410079406 A CN 202410079406A CN 117590757 B CN117590757 B CN 117590757B
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- task allocation
- population
- seagull
- 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.)
- Active
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 109
- 238000005457 optimization Methods 0.000 title claims abstract description 95
- 238000000034 method Methods 0.000 title claims abstract description 64
- 241000272168 Laridae Species 0.000 claims description 40
- 230000000739 chaotic effect Effects 0.000 claims description 40
- 238000013507 mapping Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000004891 communication Methods 0.000 claims description 11
- 238000004088 simulation Methods 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000013508 migration Methods 0.000 claims description 7
- 230000005012 migration Effects 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 239000003795 chemical substances by application Substances 0.000 claims description 6
- 238000007476 Maximum Likelihood Methods 0.000 claims description 4
- 230000001174 ascending effect Effects 0.000 claims description 3
- 238000011161 development Methods 0.000 abstract description 7
- 230000008878 coupling Effects 0.000 abstract description 3
- 238000010168 coupling process Methods 0.000 abstract description 3
- 238000005859 coupling reaction Methods 0.000 abstract description 3
- 230000006399 behavior Effects 0.000 description 15
- 230000008569 process Effects 0.000 description 15
- 230000008901 benefit Effects 0.000 description 9
- 239000002245 particle Substances 0.000 description 3
- 238000013341 scale-up Methods 0.000 description 3
- 238000012876 topography Methods 0.000 description 3
- 241000283153 Cetacea Species 0.000 description 2
- 230000009471 action Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000010845 search algorithm Methods 0.000 description 2
- 241000282461 Canis lupus Species 0.000 description 1
- 241000544061 Cuculus canorus Species 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000001617 migratory effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention provides a multi-unmanned aerial vehicle collaborative task allocation method based on a Gaussian distribution sea-gull optimization algorithm, which comprises the steps of initializing a multi-unmanned aerial vehicle collaborative combat scene, constructing a multi-unmanned aerial vehicle task allocation model, improving the sea-gull optimization algorithm based on the multi-unmanned aerial vehicle collaborative task allocation scene, solving a multi-unmanned aerial vehicle collaborative task allocation problem based on the improved sea-gull optimization algorithm, and analyzing a multi-unmanned aerial vehicle collaborative task allocation result, so that the sea-gull optimization algorithm well balances the development and exploration capacity of the algorithm by using a self-adaptive Gaussian distribution estimation strategy considering dominant population, optimal individual and self information, the optimizing performance of the algorithm is enhanced, and the tight coupling of task allocation and track planning is realized by adding a multi-unmanned aerial vehicle task allocation model of an actual track, so that the target allocation is more real and effective.
Description
Technical Field
The invention relates to a multi-unmanned aerial vehicle cooperative task allocation technology, in particular to a multi-unmanned aerial vehicle cooperative task allocation method based on a Gaussian distribution gull optimization algorithm.
Background
The task allocation of multiple unmanned aerial vehicles is an important content of unmanned aerial vehicle task planning, namely, in a complex battlefield environment, ordered tasks are reasonably allocated for unmanned aerial vehicles on own, so that the overall combat effectiveness is maximized. The problem of multi-unmanned aerial vehicle task allocation is essentially a multi-constraint combined optimization problem, and is a typical 'NP-hard' problem. The core of the problem mainly comprises two aspects of task allocation model establishment and model solving algorithm.
In the aspect of task allocation model establishment, the task allocation model commonly used at home and abroad comprises: vehicle path problem model, multi-traveller problem model, contract net auction model, mixed integer linear programming model, etc. Task allocation models are generally classified into a single task allocation model and a multi-task allocation model according to task complexity. The single-mission model generally employs a vehicle path problem model and a multi-traveler problem model. As the types of executing tasks increase, the single-task allocation model cannot meet the task planning requirements, and thus a multitasking model is proposed to meet the task planning requirements. The multitasking distribution model generally adopts a mixed integer linear programming model, a dynamic network flow optimization model and the like.
In terms of model solving algorithms, there are generally two categories: one class is traditional optimization methods such as dynamic programming, mixed integer linear programming. The other class is intelligent optimization algorithm, such as whale optimization algorithm, particle swarm optimization algorithm, genetic algorithm. Along with the increasing of the complexity of the problem, the traditional optimization method is difficult to obtain a high-quality solution in a reasonable time, the group intelligent algorithm does not depend on a problem model, gradient information is not needed, and the method has the advantages of being strong in searching capability, wide in application range and the like, and is widely used for solving the unmanned aerial vehicle task planning problem.
The gull optimization algorithm (SOA, seagull optimization algorithm) is an optimization algorithm inspired by animal behaviors in Gaurav Dhiman in 2018, and the inspiration mainly comes from the migratory behaviors and the attack behaviors of the gull. SOA has better performance than genetic algorithms, particle swarm algorithms, gravitational search algorithms, cuckoo search algorithms, goblet-sea squirt algorithms, differential evolution algorithms, and thus has been widely used to address a number of practical engineering problems, such as: photovoltaic field, power system, task scheduling. However, the SOA has some disadvantages, and when the SOA updates the population position, the SOA mainly searches near the optimal individuals, but does not utilize the effective information of more individuals, which easily causes the population diversity to be reduced and easily falls into local optimization. Furthermore, SOAs are often used for continuous optimization problems, which are difficult to solve for discrete problems.
The conventional task allocation model generally uses straight line distance between an unmanned plane and a target when considering the distance between the unmanned plane and the target, but in the actual combat process, the actual path of the unmanned plane is not a straight line but a curve, compared with the straight line, due to the need of avoiding reconnaissance and threat and following the terrain flight, the distance is increased, which may result in that the task allocation result with the maximum actual combat effectiveness cannot be obtained, and meanwhile, most task allocation models only can give the target hit sequence and cannot plan the specific flight path due to the fact that only the straight line distance is considered.
Disclosure of Invention
Aiming at the problems, the invention provides a multi-unmanned aerial vehicle collaborative task allocation method based on a Gaussian distribution sea-gull optimization algorithm.
The technical scheme adopted is that the multi-unmanned aerial vehicle collaborative task allocation method based on Gaussian distribution sea-gull optimization algorithm comprises the following steps:
S1, initializing a multi-unmanned aerial vehicle collaborative combat scene;
S2, constructing a task allocation model of the multiple unmanned aerial vehicles;
S3, improving a seagull optimization algorithm based on a multi-unmanned aerial vehicle cooperative task allocation scene;
S4, solving a multi-unmanned aerial vehicle collaborative task allocation problem based on an improved seagull optimization algorithm;
s5, analyzing the result of the multi-unmanned aerial vehicle collaborative task allocation.
Further, in S1, let the unmanned aerial vehicle set be u= { U 1,U2,...,UNU }, NU be the number of unmanned aerial vehicles;
The hit task target set is t= { T 1,T2,...,TNT }, NT is the target number.
Optionally, in S2, the multi-unmanned aerial vehicle task allocation model structure is performed from five dimensions, including the task allocation model benefit, the flight distance matrix, the target allocation constraint condition, the communication distance constraint condition and the single-machine total range constraint condition, respectively.
Further, in S2, the function constructed according to the task allocation model profit is:
Wherein,
I i represents an attack target sequence distributed by the ith unmanned aerial vehicle;
η (I i (k)) represents the threat value of the I i (k) th target;
D UT(i,Ii (1)) represents the actual flight distance of the current drone from the first attack target;
D TT(Ii(k-1),Ii (k)) represents the actual flight distance of the kth-1 target from the kth target;
Representing the normalized distance;
the function constructed from the flight distance matrix is:
the function constructed from the target allocation constraints is:
Wherein,
X ij represents the allocation result of the jth target to the ith unmanned aerial vehicle, a value of 1 represents allocation, and a value of 0 represents no allocation;
I max is the maximum value of the number of target hits performed by the unmanned aerial vehicle;
the function constructed according to the communication distance constraint condition is as follows:
Wherein,
Representing the distance between the ith unmanned aerial vehicle and the ground command system;
D Cmax represents the maximum communication radius of the ground control system;
the function constructed according to the single machine total range constraint condition is:
Wherein,
I i is an attack target sequence of the ith unmanned aerial vehicle;
i i is the total number of attack targets allocated to the ith unmanned aerial vehicle;
D UT(i,Ii (1)) is the range of the ith unmanned aerial vehicle from the first target;
D TT(Ii(k),Ii (k+1)) is the distance between the kth target position and the kth+1th target position of sequence I i;
meanwhile, the total range constraint requires that the total range required by each unmanned aerial vehicle to implement attack satisfies the following functions:
Wherein,
D max is the maximum range of the unmanned aerial vehicle.
Optionally, constructing a multi-unmanned aerial vehicle task allocation model based on a function constructed in five dimensions, wherein the objective function is as follows:
J＝R-λ·(gA+gC+gD)
Wherein,
Lambda is penalty factor, and the value is 10 5;
The larger J indicates the better the yield of the task allocation result.
Further, in S3, the method includes the following steps:
A1. initializing a seagull population through an Arnold chaotic mapping strategy;
A2. Calculating the fitness value of the individual seagull optimization algorithm;
A3. the seagull optimization algorithm is improved by a stagnation judging strategy to judge whether the seagull optimization algorithm falls into local optimization;
A4. Updating the gull population position in the gull optimization algorithm;
A5. And judging whether the iteration ending condition is met.
Optionally, in A1, the kinetic equation calculation formula of the Arnold chaotic strategy is:
Wherein,
Mod represents a modulo operation;
The chaotic sequence interval generated by Arnold chaotic strategy is [0,1];
h n and k n represent the current values of the chaotic sequences h and k, respectively;
h n+1 and k n+1 represent the next time values of the chaotic sequences h and k, respectively;
the chaotic sequence generated by Arnold chaotic mapping strategy initializes the standard seagull optimization algorithm population, and the method is specifically as follows:
Xi,j＝lbj+h·(ubj-lbj)
Wherein,
X i,j represents the j (j=1, 2, …, D) dimension of the i (i=1, 2, …, N) th agent in the initial population;
lb j and ub j represent the upper and lower bound ranges of the j-th dimension of the vector;
In A2, the fitness value calculation formula is:
fitnessi＝f(Xi)
Wherein,
Fitness i represents the fitness value of the i (i=1, 2, …, N) th agent;
f (Xi) represents a multi-unmanned task allocation objective function of the individual X i;
In the A3, the stagnation judging strategy judges whether the local optimization is trapped or not by comparing whether the optimal average positions of the front population and the back population are the same, and the judging strategy is as follows:
Wherein,
F stag is a flag indicating whether the population falls into a local optimum;
X mean represents the individual optimal average position;
X lbest,i represents the current optimal position of the ith seagull;
And A4, according to the judging result of the stagnation judging strategy, updating the gull population position by adopting two strategies:
strategy 1, when F stag =0, shows that the seagull population is not trapped into local optimum, and the individual seagull population updates the position information through the migration behavior and the attack behavior of the standard seagull optimization algorithm;
Strategy 2, when F stag =1, representing that the seagull population falls into local optimum, taking the first half population with the minimum fitness value as the dominant population, calculating a probability distribution model through the dominant population of the current iteration times and a weighted maximum likelihood estimation method, and generating a new offspring population according to the probability distribution model;
In A5, judging whether the termination condition is met, outputting an optimal task allocation scheme, if the iteration number reaches the maximum iteration number, ending the iteration, outputting an optimal solution, and decoding the optimal solution into a multi-unmanned aerial vehicle task allocation scheme; otherwise, continue A2.
Further, in S4, the method includes the following steps:
B1. initializing a multi-unmanned aerial vehicle collaborative task allocation scene;
B2. Establishing a mapping relation between an individual position and a task allocation result through a coding mode based on a real number vector;
B3. determining the number of gull population and the maximum iteration number of algorithm optimization;
B4. Calculating a fitness value function;
B5. updating the gull population position;
B6. And judging whether a termination condition is reached.
Optionally, in B1, it is assumed that an NU frame unmanned aerial vehicle performs an attack task in a known airspace, and NT targets in a task area need to be hit;
In the step B2, the optimal dimension of the multi-unmanned aerial vehicle collaborative task allocation problem is the target number required to be hit, the dimension number of an individual corresponds to the target number, the search space range of the solution is (1, N U +1), the integral part of the individual position corresponds to the number of the UAV, and the decimal part corresponds to the order of executing the task of the same UAV in ascending order;
In the step B3, initializing a seagull population according to a coding mode based on a real number vector, generating a uniformly distributed chaotic sequence in a [0,1] interval by adopting an Arnold chaotic mapping strategy and searching the initialized seagull population by utilizing the chaotic sequence generated by the Arnold chaotic mapping strategy;
in the step B4, decoding the individual position information of the population into a task allocation mapping relation, and converting the multi-constraint multi-unmanned aerial vehicle task allocation problem into an unconstrained optimization problem by adopting a penalty function method:
fitness＝maxJ
the task allocation problem of the multiple unmanned aerial vehicles is a maximized problem, and the larger the objective function value of the model is, the larger the fitness value of the population individuals is, and the better the task allocation solution is;
in B5, whether the sea gull population is trapped in local optimum is judged by comparing whether the optimal average positions of the sea gull population before and after are the same, when F stag =0, the sea gull population is not trapped in local optimum, policy 1 is adopted as a stagnation judgment policy, when F stag =0, the sea gull population is trapped in local optimum, and policy 2 is adopted as a stagnation judgment policy;
In step B6, judging whether the termination condition is met, outputting an optimal task allocation scheme, if the iteration number reaches the maximum iteration number, ending the iteration, outputting an optimal solution, and decoding the optimal solution into a multi-unmanned aerial vehicle task allocation scheme; otherwise, continue B4.
Further, in S5, the method includes the following steps:
C1. Comparing and simulating different distance models;
C2. task allocation model feasibility simulation;
C3. and (5) expanding the task scale simulation.
The beneficial effects of the invention at least comprise one of the following;
1. By adding the multi-unmanned aerial vehicle task allocation model of the actual flight path, the tight coupling of task allocation and flight path planning is realized, so that the target allocation is more real and effective.
2. By using a self-adaptive Gaussian distribution estimation strategy considering dominant population, optimal individual and self information, the seagull optimization algorithm well balances the development and exploration capabilities of the algorithm and enhances the optimizing performance of the algorithm.
3. The method for encoding based on the real number vector successfully applies the seagull optimization algorithm to the mixed integer programming problem.
4. The problem that the existing task allocation model generally uses straight line distance between an unmanned aerial vehicle and a target when considering the distance between the unmanned aerial vehicle and the target, but in the actual combat process, the actual path of the unmanned aerial vehicle is not a straight line but a curve, compared with the straight line, the distance is increased due to the fact that the actual path of the unmanned aerial vehicle is not a straight line, and the maximum actual combat efficiency task allocation result cannot be obtained, and meanwhile, most task allocation models only can give out target hitting sequence and cannot plan a specific flight path due to the fact that only the straight line distance is considered.
Drawings
FIG. 1 shows the improvement strategy of the seagull optimization algorithm;
FIG. 2 shows a flow chart for improving the SOA algorithm;
FIG. 3 shows a flow chart of a multi-unmanned aerial vehicle collaborative task allocation method based on an improved gull optimization algorithm;
FIG. 4 illustrates a task execution trace graph for a scale-up operating condition.
Detailed Description
Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention. It should be noted that the following embodiments and features in the embodiments may be combined with each other without conflict.
It should be noted that the illustrations provided in the following embodiments merely illustrate the basic concept of the present invention by way of illustration, and only the components related to the present invention are shown in the drawings and are not drawn according to the number, shape and size of the components in actual implementation, and the form, number and proportion of the components in actual implementation may be arbitrarily changed, and the layout of the components may be more complicated.
Example 1
A multi-unmanned aerial vehicle cooperative task allocation method based on a Gaussian distribution seagull optimization algorithm comprises the following steps:
S1, initializing a multi-unmanned aerial vehicle collaborative combat scene;
S2, constructing a task allocation model of the multiple unmanned aerial vehicles;
S3, improving a seagull optimization algorithm based on a multi-unmanned aerial vehicle cooperative task allocation scene;
S4, solving a multi-unmanned aerial vehicle collaborative task allocation problem based on an improved seagull optimization algorithm;
s5, analyzing the result of the multi-unmanned aerial vehicle collaborative task allocation.
The design aims at realizing the tight coupling of task allocation and track planning by adding the multi-unmanned aerial vehicle task allocation model of the actual track, so that the target allocation is more real and effective. By using a self-adaptive Gaussian distribution estimation strategy considering dominant population, optimal individual and self information, the seagull optimization algorithm well balances the development and exploration capabilities of the algorithm and enhances the optimizing performance of the algorithm. The method for encoding based on the real number vector successfully applies the seagull optimization algorithm to the mixed integer programming problem. The problem that the existing task allocation model generally uses straight line distance between an unmanned aerial vehicle and a target when considering the distance between the unmanned aerial vehicle and the target, but in the actual combat process, the actual path of the unmanned aerial vehicle is not a straight line but a curve, compared with the straight line, the distance is increased due to the fact that the actual path of the unmanned aerial vehicle is not a straight line, and the maximum actual combat efficiency task allocation result cannot be obtained, and meanwhile, most task allocation models only can give out target hitting sequence and cannot plan a specific flight path due to the fact that only the straight line distance is considered.
Example 2
A multi-unmanned aerial vehicle cooperative task allocation method based on a Gaussian distribution seagull optimization algorithm comprises the following steps:
S1, initializing a multi-unmanned aerial vehicle collaborative combat scene;
S2, constructing a task allocation model of the multiple unmanned aerial vehicles;
S3, improving a seagull optimization algorithm based on a multi-unmanned aerial vehicle cooperative task allocation scene;
S4, solving a multi-unmanned aerial vehicle collaborative task allocation problem based on an improved seagull optimization algorithm;
s5, analyzing the result of the multi-unmanned aerial vehicle collaborative task allocation.
In S1, it is assumed that a plurality of unmanned aerial vehicles perform an attack task in a known airspace, and a target in a task area needs to be hit. Let the unmanned aerial vehicle set be U= { U 1,U2,...,UNU }, NU is unmanned aerial vehicle number; the task target set is T= { T 1,T2,...,TNT }, NT is the target number, and the fight principle of the unmanned aerial vehicle is to eliminate the high threat target as soon as possible, so that the threat degree and attack range of the target are two factors to be considered in task allocation research.
In S2, the construction of the multi-unmanned aerial vehicle allocation model needs to be considered from two aspects: firstly, an objective function; and secondly, constraint conditions. The objective function is generally related to the benefits of the task allocation model, and the greater the benefits of the combat task, the more beneficial the allocation scheme to combat, and the higher the combat effectiveness of the unmanned aerial vehicle. The construction of the mission allocation benefit function is not only related to the threat of the target, but also related to the track path for executing the mission. In order to improve the authenticity of task allocation, the track cost of executing the task needs to consider factors such as terrain, threat and the like, so the invention adopts the actual flight distance as a part of the profit function. In order to meet the objectivity of the model, constraint conditions of the multi-unmanned aerial vehicle collaborative task allocation model comprise target allocation constraint, communication distance constraint and single machine total range constraint.
In the mission allocation model revenue function construction, the revenue J i of the ith unmanned attack on the jth target is related to the threat value T j of the target and the track cost of performing the mission. Based on the method, the integral hitting income function of the laser attack unmanned aerial vehicle is constructed as follows:
wherein I i represents an attack target sequence allocated by the ith unmanned aerial vehicle, η (I i (k)) represents a threat value of the I i (k) th target; d UT(i,Ii (1)) represents the actual flight distance of the current drone from the first attack target, D TT(Ii(k-1),Ii (k)) represents the actual flight distance of the kth-1 target from the kth target, and D represents the normalized distance. From the formula, the different time sequences of the distributed attack targets can lead to different required attack distances, and finally the overall attack benefits are influenced.
In the flight distance matrix construction, the distances between the unmanned aerial vehicle and the target and between the targets in most task allocation models are all Euclidean distances used. However, in the actual combat process, the actual flight path of the unmanned aerial vehicle is not necessarily a straight line due to factors such as topography, threat and the like, and the unmanned aerial vehicle can fly and avoid the threat to perform curve motion due to low-altitude ground-attached flight, so that the Euclidean distance is used as the flight distance in the task allocation model, and the flight distance is not fit with the actual situation. In order to fully consider three-dimensional terrain information and threat factors, the invention uses an unmanned aerial vehicle three-dimensional track planning model as flight distance input, and constructs a distance matrix to store the flight distances of the UAVs and the targets, the attack targets and the attack targets in order to avoid repeated calculation in the optimization process.
In the target allocation constraint condition construction, the target allocation result needs to ensure that each target is allocated with a corresponding laser unmanned aerial vehicle to strike, and meanwhile, the maximum attack capacity of each unmanned aerial vehicle meets the weapon load constraint. The constraint conditions are as follows:
Wherein x ij represents the allocation result of the jth target to the ith unmanned aerial vehicle, a value of 1 represents allocation, a value of 0 represents no allocation, and I max is the maximum value of the number of target hitting performed by the unmanned aerial vehicle. The formula indicates that each target can only be allocated to one unmanned aerial vehicle, and all targets are allocated; and meanwhile, considering the limitation of the airborne weapon, the number of targets allocated to each unmanned aerial vehicle cannot exceed the maximum executable task number.
In the communication distance constraint condition construction, in order to obtain accurate target situation information, the laser unmanned aerial vehicle needs to execute tasks in the communication coverage range of the ground control system, so that the communication distance constraint is constructed as follows:
Wherein, The distance between the ith unmanned aerial vehicle and the ground control system is represented, and D Cmax represents the maximum communication radius of the ground control system.
In a stand-alone total range constraint configuration, the total range D i for the ith unmanned aerial vehicle to perform all assigned hit tasks can be expressed as:
Wherein I i is a hit target sequence of the ith unmanned aerial vehicle, I i is the total number of attack targets distributed by the ith unmanned aerial vehicle, D UT(i,Ii (1)) is the range of the ith unmanned aerial vehicle from the first target, and D TT(Ii(k),Ii (k+1)) is the distance between the kth target position and the k+1 target position of the sequence I i.
The total range constraint requires that the total range required by each unmanned aerial vehicle to implement the attack meets the design requirement, namely:
Wherein D max is the maximum range of the unmanned aerial vehicle.
In the objective function construction, the above benefit function and constraint conditions are considered, a penalty function method is adopted to convert the multi-constraint optimization problem into an unconstrained optimization problem, and the objective function is constructed as follows:
J＝R-λ·(gA+gC+gD)
Wherein lambda is a penalty factor and takes on a value of 10 5. The larger J indicates the better the yield of the task allocation result.
In S3, as shown in fig. 1 and fig. 2, the method for solving the multi-unmanned aerial vehicle cooperative task allocation problem includes a conventional optimization method and a group intelligent optimization algorithm. When solving the high-dimensional optimization problem, the traditional optimization method is difficult to obtain a high-quality solution and is easy to sink into local optimization. The group intelligent optimization algorithm does not depend on gradient information when solving the complex optimization problem, has stronger searching capability, and is widely applied to the unmanned aerial vehicle task planning problem. The Seagull Optimization Algorithm (SOA) is used as a novel natural heuristic intelligent optimization algorithm, and is mainly applied to the research fields of path planning, bridge structure optimization, photovoltaic device diagnosis, control parameter optimization and the like. The seagull optimization algorithm has the advantages of simple structure, few calculation parameters and easy operation, but has lower calculation precision and is easy to generate the premature phenomenon when solving the complex problem. In order to overcome the defect of optimization precocity of classical seagulls, the solving precision of the multi-unmanned aerial vehicle cooperative task allocation problem is improved, and the convergence speed of the optimization process is accelerated.
Therefore, the Arnold chaotic mapping strategy is utilized to initialize the seagull population in A1. In a standard sea-gull optimization algorithm (SOA), population positions generated in a population initialization mode are unevenly distributed and poor in stability, so that the precision of search optimization is reduced. Arnold chaotic strategy is a two-dimensional reversible chaotic mapping mode, has a simple structure, and can generate uniformly distributed chaotic sequences in the [0,1] interval. In order to increase initial population diversity, the invention adopts Arnold chaotic strategy to improve SOA algorithm. The kinetic equation of the Arnold chaotic strategy is calculated as follows:
where mod represents a modulo operation;
The chaotic sequence interval generated by Arnold chaotic strategy is [0,1];
h n and k n represent the current values of the chaotic sequences h and k, respectively;
h n+1 and k n+1 represent the next time values of the chaotic sequences h and k, respectively.
The chaotic sequence generated by Arnold chaotic mapping strategy is used for initializing the SOA population, and the method is specifically as follows:
Xi,j＝lbj+h·(ubj-lbj)
Where X i,j represents the j (j=1, 2, …, D) dimension of the i (i=1, 2, …, N) th agent in the initial population, lb j and ub j represent the upper and lower bound ranges of the j (th) dimension of the vector.
And then, calculating the fitness value of the SOA algorithm individual in A2, wherein the specific calculation formula is as follows:
fitnessi＝f(Xi)
Where fitness i represents the fitness value of the i (i=1, 2, …, N) th agent, and f (Xi) represents the multiple unmanned aerial vehicle task allocation objective function of individual X i.
Next, in A3, in order to overcome the problem of trapping in local optimum, the present invention adopts a stagnation judgment strategy to improve the SOA algorithm. The stagnation judging strategy judges whether the local optimization is trapped or not by comparing whether the optimal average positions of the front population and the back population are the same. The specific judgment strategy is as follows:
wherein, F stag represents a flag of whether the population falls into a local optimum, X mean represents an individual optimum average position, and X lbest,i represents an ith seagull current optimum position.
Then, in A4, the sea gull population position of the SOA algorithm is updated. And according to the judging result of the stagnation judging strategy, the SOA algorithm adopts different strategies to update the positions of the seagull population. F stag =0 indicates that the seagull population is not trapped in local optimum, and the invention adopts a standard SOA algorithm seagull updating strategy (strategy 1); f stag =1 indicates that the gull population is trapped in a local optimum, and the invention adopts a gaussian distribution estimation strategy (strategy 2).
Strategy 1: standard SOA algorithm seagull updating strategy. When F stag =0, the seagull population is not trapped in local optimum, and the invention still adopts a standard SOA algorithm seagull updating method. The standard SOA algorithm sea-gull updating method comprises a migration behavior and an attack behavior. The migration behavior is the behavior that the seagull approaches to the optimal seagull position on the premise of avoiding collision, and the hunting object is searched in the whole space, and the exploration process of an algorithm is mainly simulated. The process of approaching the prey by the action of spiral moving and flying is mainly the exploitation process of simulation algorithm.
(1) Migration behaviour
In the migration process, the position update of the seagull population mainly meets three conditions: avoiding collision between seagulls, approaching the optimal seagull individuals, and updating the position of the seagull individuals.
In order to avoid collision between two seagulls, the positions of individual seagulls after movement are calculated by using a variable A:
Cs＝A×Xs(t)
Wherein, C s represents the position of the seagull searching individual which does not collide with other individuals, X s represents the current position of the seagull individual, t represents the iteration times, and the variable A represents the movement behavior of the individual, and the specific calculation is as follows:
A＝fc-(t·fc/Maxiteration)
where f c is the number of times the manipulated variable A is used, and is generally set to 2.Max iteration represents the maximum number of iterations. The value of the variable A decreases from 2 to 0 with the increase of the iteration times.
After avoiding collisions between adjacent individuals, the searching individual will move toward the optimal individual as follows:
Ms＝B×(Xbs-Xs)
Where M s represents the new position after the individual gull has moved to the optimal individual gull X bs. B is a random number, which is used mainly to balance the exploration and development capabilities of the algorithm, and is calculated as follows:
B＝2·A2·rand
wherein, rand is a random number, and the value range is between 0 and 1.
Finally, the search individual position movement distance formula is as follows with respect to the optimal individual:
Ds＝|Cs+Ms|
Where D s represents the distance between the searching individual and the optimal individual.
(2) Attack behavior
The development process of the attack behavior simulation optimization algorithm mainly utilizes the history experience in the searching process. The attack direction is mastered by the wings and weight in the process of attacking the prey by the seagull. In general, the process of seagull attack hunting is generally simulated by adopting spiral movement, and a specific calculation formula is as follows:
Xs(t+1)＝(Ds×x×y×z)+Xbs(t)
x＝r×cos(k)
y＝r×sin(k)
z＝r×k
r＝u×ekv
Wherein X s (t+1) represents the gull individual after the position update. r represents the radius of rotation of the spiral and k represents a random number between 0 and 2 pi. u and v represent constants defining the spiral motion, e being the base of the natural logarithm. The seagull optimization algorithm adopts a random method to generate a population, and updates the position of the seagull population by converting the global searching and the local searching processes, so that the optimal individual is finally searched.
Strategy 2: gaussian distribution estimation strategy. If F stag =1, the gull population is trapped in a local optimum, the invention still adopts a gaussian distribution estimation strategy to update the position of the gull population. Gaussian distribution estimation strategies represent inter-individual relationships by probabilistic models. The method utilizes the current dominant population to calculate a probability distribution model, and generates a new offspring population according to the probability distribution model, so that the optimal solution is finally obtained through continuous iteration. The invention utilizes a weighted maximum likelihood estimation method to estimate a distribution model, takes the first half population with better performance as a dominant population, and the mathematical model of the algorithm is described as follows:
Xi＝mean+y,y～N(0,Cov)
mean＝0.7(w1·Xmean+w2·Elite)+0.3Xi
w2＝1-w1
Wherein X mean represents the weighted average of the dominant population, NP represents the population number, omega i represents the weight coefficients of the dominant population arranged in descending order of fitness values, and Cov is the weighted covariance matrix of the dominant population. w 1 and w 2 respectively represent the weighted mean and the influence weight of the globally optimal individual, and in the early stage of optimization, the influence weight of the weighted mean is larger because the influence weight is explored as large as possible; and local development is performed as much as possible in the later stage, so that the optimal individual influence weight is larger. Besides the introduction of social information, the strategy also introduces individual information, so that the waste of the individual information is avoided. Therefore, under the three actions of weighted average, optimal individuals and self, the individuals can consider more the population evolution trend and self information in the early stage, and consider more the optimal individuals and self information in the later stage, so that the development and the exploration of the algorithm are balanced, and the key point of the intelligent optimization algorithm in improving the performance is the key point of the intelligent optimization algorithm.
Finally, in A5, it is determined whether the iteration end condition is satisfied. If the iteration number iter reaches the maximum iteration number iter max, ending the iteration and outputting an optimal value; otherwise, continuously calculating the adaptability value of the SOA algorithm individual.
In S4, in order to search a reasonable task allocation scheme, the invention provides a multi-unmanned aerial vehicle collaborative task allocation method based on an improved seagull optimization algorithm. The key points of the multi-unmanned aerial vehicle cooperative task allocation method mainly comprise a task allocation coding mode and an adaptability function structure. The specific flow of the multi-unmanned aerial vehicle cooperative task allocation method based on the improved seagull optimization algorithm is shown in fig. 3, and the method comprises the following steps:
B1. initializing a multi-unmanned aerial vehicle collaborative task allocation scene;
B2. Establishing a mapping relation between an individual position and a task allocation result through a coding mode based on a real number vector;
B3. determining the number of gull population and the maximum iteration number of algorithm optimization;
B4. Calculating a fitness value function;
B5. updating the gull population position;
B6. And judging whether a termination condition is reached.
In B1, initializing a multi-unmanned aerial vehicle cooperative task allocation scene. Assuming that an NU-frame unmanned aerial vehicle executes an attack task in a known airspace, NT targets in a task area need to be hit. The task space is 100km×100km, the coordinate positions of the unmanned aerial vehicle and the target are initialized, and specific information is shown in table 1:
TABLE 1
And B2, establishing a mapping relation between the individual position and the task allocation result by using a coding mode based on a real number vector. For convenience in describing the mapping relationship, the present invention makes the following definitions:
(1) The target number of the striking required by the optimization dimension of the multi-unmanned aerial vehicle cooperative task allocation problem is corresponding to the individual dimension number;
(2) The search space range of the solution is (1, N U +1);
(3) The individual position integer part corresponds to the serial number of the UAV, and the decimal part corresponds to the order of executing tasks of the same UAV in ascending order. For a clearer description of the mapping relationship, table 2 shows:
TABLE 2
The 3 UAV attacks on 7 targets are illustrated as examples. The optimization dimension of the current problem is 7, and the search range of the particle optimizing space is (1, 4).
And B3, determining the number of gull population and optimizing the maximum iteration times by an algorithm. And initializing the seagull population according to the coding mode based on the real number vector. The seagull search space, namely the search space of the task allocation scheme, is abstracted into European space, and the Arnold chaotic mapping strategy is adopted to generate a uniformly distributed chaotic sequence in the [0,1] interval. And searching and initializing a seagull population by using a chaotic sequence generated by Arnold chaotic mapping strategy.
In B4, a fitness value function is calculated. Firstly, decoding the group individual position information into a task allocation mapping relation, namely, an unmanned aerial vehicle hitting target time sequence number for calculating a multi-unmanned aerial vehicle task allocation model objective function. Then, according to the objective function of the multi-unmanned aerial vehicle task allocation model established in the step B2, simultaneously converting the multi-constraint multi-unmanned aerial vehicle task allocation problem into an unconstrained optimization problem by adopting a penalty function method:
fitness＝maxJ
the task allocation problem of the multiple unmanned aerial vehicles is a maximized problem, and the larger the objective function value of the model is, the larger the fitness value of the population individuals is, and the better the task allocation solution is.
And B5, updating the gull population position. Firstly, adopting a stagnation judging strategy, and judging whether the sea gull population is in local optimum or not by comparing whether the optimal average positions of the sea gull population are the same or not.
If F stag =0, it means that the gull population is not trapped in local optimum, the invention adopts standard SOA algorithm gull update strategy (strategy 1). In strategy 1, the gull population individuals update the position information through the migration behavior and the attack behavior of the SOA algorithm.
If F stag =0, which indicates that the gull population is trapped locally optimally, the invention adopts gaussian distribution estimation strategy (strategy 2). Firstly, taking the first half population with the smallest fitness value as a dominant population, and calculating a probability distribution model by using the dominant population of the current iteration times and a weighted maximum likelihood estimation method; the generation of new offspring populations is then employed according to the probability distribution model.
And B6, judging whether the termination condition is met, and outputting an optimal task allocation scheme. If the iteration number reaches the maximum iteration number, finishing the iteration, outputting an optimal solution, and decoding the optimal solution into a multi-unmanned aerial vehicle task allocation scheme; otherwise, continue B4.
In S5, the multi-unmanned aerial vehicle cooperates with the task allocation result analysis. According to the set multi-unmanned aerial vehicle cooperative task allocation scene, the invention carries out simulation analysis from three aspects of distance model comparison, task allocation model feasibility, task scale expansion and the like, and comprises the following steps:
C1. Comparing and simulating different distance models;
C2. task allocation model feasibility simulation;
C3. and (5) expanding the task scale simulation.
In C1, different distance models compare simulation analysis. Most task allocation models use the straight line distance between two points when calculating the distance, and neglect the influence of topography factors, so that the allocation result is not practical. In order to verify the validity and necessity of the mission allocation model based on the flight distance of the invention, according to the A2 model of S2, the pre-planned flight distance is used as the mission allocation model distance as input. The experiment does not consider threat factors, only considers topography factors, and the linear distances are shown in table 3:
TABLE 3 Table 3
The flight distance considering only the distance cost is shown in table 4:
TABLE 4 Table 4
The flight distance considering the distance cost and the altitude cost is shown in table 5:
TABLE 5
Based on the three distance input methods, solving is performed respectively, and respective distance matrixes are counted.
Combining the distance matrix, calculating an objective function, obtaining an optimal task distribution result through improved SOA algorithm iteration, and comparing and analyzing optimal task distribution modes of different distance models, wherein the distribution result is shown in a table 6:
TABLE 6
In C2, task allocation model feasibility simulation analysis. And solving the proposed task allocation model by utilizing an improved SOA algorithm, obtaining an objective function value of the task allocation model, and analyzing the optimal task execution sequence. In order to show the effectiveness and better optimizing capability of an Improved SOA algorithm (ISOA),
Comparing and analyzing with SOA algorithm, whale optimization algorithm (WOA, whaleoptimizationalgorithm), gray wolf optimization algorithm (GEDGWO, greywolfoptimizerbasedon Gaussian estimation distribution) based on Gaussian distribution model and ISOA algorithm, running ten times, counting average value, standard deviation and maximum value of task allocation objective function, wherein the average value of task allocation objective function represents the whole level of the proposed algorithm, the standard deviation of objective function represents the discrete degree of the objective function value of the proposed algorithm, the maximum value of objective function represents the optimal solution level of the proposed algorithm, and specific allocation comparison results are shown in tables 7 and 8:
TABLE 7
TABLE 8
In C3, the task scale simulation analysis is expanded. In order to verify the robustness of the improved SOA algorithm to solve the multi-unmanned aerial vehicle task allocation problem, the embodiment utilizes the improved SOA algorithm to solve a large-scale multi-unmanned aerial vehicle task allocation model. The method is characterized in that the method is expanded to 4 unmanned aerial vehicles to strike 16 targets on the basis of striking 9 targets by the 3 unmanned aerial vehicles, and the expanded part of unmanned aerial vehicles and target information are shown in table 9:
TABLE 9
And solving by adopting an improved SOA algorithm, running for ten times, and counting the average value, standard deviation and maximum value of the task allocation objective function.
The optimal task allocation results are obtained through calculation of objective functions and optimization iteration, and the optimal task allocation schemes of different task scales are compared and analyzed, wherein the task allocation statistical results of the scale expansion working conditions are shown in the table 10:
Table 10
The best task execution sequence for the scale-up condition is shown in table 11:
TABLE 11
The task execution trace under the scale-up condition is shown in fig. 4.
Finally, it should be noted that: the foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, it will be apparent to those skilled in the art that the foregoing description of the preferred embodiments of the present invention can be modified or equivalents can be substituted for some of the features thereof, and any modification, equivalent substitution, improvement or the like that is within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (4)
1. The multi-unmanned aerial vehicle collaborative task allocation method based on the Gaussian distribution seagull optimization algorithm is characterized by comprising the following steps of:
S1, initializing a multi-unmanned aerial vehicle collaborative combat scene;
let the unmanned aerial vehicle set be U= { U 1,U2,...,UNU }, NU is unmanned aerial vehicle number;
The target set of the hitting task is T= { T 1,T2,...,TNT }, and NT is the target number;
S2, constructing a multi-unmanned aerial vehicle task allocation model taking track planning into consideration; constructing a multi-unmanned aerial vehicle task allocation model from five dimensions in total according to the task allocation model profit D1, the flight distance matrix D2, the target allocation constraint condition D3, the communication distance constraint condition D4 and the single-machine total range constraint condition D5; the multi-unmanned aerial vehicle task allocation model construct includes the following dimensions:
D1. The function of the revenue construction according to the task allocation model is:
Wherein,
I i represents an attack target sequence distributed by the ith unmanned aerial vehicle;
η (I i (k)) represents the threat value of the I i (k) th target;
D UT(i,Ii (1)) represents the actual flight distance of the current drone from the first attack target;
D TT(Ii(k-1),Ii (k)) represents the actual flight distance of the kth-1 target from the kth target;
Representing the normalized distance;
D2. the function constructed from the flight distance matrix is:
Wherein,
D UT (NU, NT) represents the actual flight distance of the NU-th frame drone from the NT-th attack target;
D TT (NT, NT) represents the actual flight distance of the NT-th attack target from the NT-th attack target;
Distance max represents the maximum flight Distance in the flight Distance matrix;
D3. The function constructed from the target allocation constraints is:
Wherein,
X ij represents the allocation result of the jth target to the ith unmanned aerial vehicle, a value of 1 represents allocation, and a value of 0 represents no allocation;
I max is the maximum value of the number of target hits performed by the unmanned aerial vehicle;
D4. the function constructed according to the communication distance constraint condition is as follows:
Wherein,
Representing the distance between the ith unmanned aerial vehicle and the ground command system;
D Cmax represents the maximum communication radius of the ground control system;
D5. The function constructed according to the single machine total range constraint condition is:
Wherein,
I i is an attack target sequence of the ith unmanned aerial vehicle;
i i is the total number of attack targets allocated to the ith unmanned aerial vehicle;
D UT(i,Ii (1)) is the range of the ith unmanned aerial vehicle from the first target;
D TT(Ii(k),Ii (k+1)) is the distance between the kth target position and the kth+1th target position of sequence I i;
meanwhile, the total range constraint requires that the total range required by each unmanned aerial vehicle to implement attack satisfies the following functions:
Wherein,
D max is the maximum range of the unmanned aerial vehicle;
And based on the functions of the five-dimensional construction, the multi-unmanned aerial vehicle task allocation model construction is carried out, and the objective functions are as follows:
J＝R-λ·(gA+gC+gD)
Wherein,
Lambda is penalty factor, and the value is 10 5;
The larger J is, the better the income of the task allocation result is;
S3, improving a seagull optimization algorithm based on a multi-unmanned aerial vehicle cooperative task allocation scene; the improved seagull optimization algorithm comprises the following steps:
A1. Initializing a seagull population through an Arnold chaotic mapping strategy; the calculation formula of the dynamics equation of the Arnold chaotic strategy is as follows:
Wherein,
Mod represents a modulo operation;
The chaotic sequence interval generated by Arnold chaotic strategy is [0,1];
h n and k n represent the current values of the chaotic sequences h and k, respectively;
h n+1 and k n+1 represent the next time values of the chaotic sequences h and k, respectively;
the chaotic sequence generated by Arnold chaotic mapping strategy initializes the standard seagull optimization algorithm population, and the method is specifically as follows:
Xi,j＝lbj+h·(ubj-lbj)
Wherein,
X i,j represents the j-th dimension of the i-th agent in the initial population, and i=1, 2, …, N, j=1, 2, …, D;
lb j and ub j represent the upper and lower bound ranges of the j-th dimension of the vector;
A2. Calculating the fitness value of the individual seagull optimization algorithm; according to the description and analysis of the multi-unmanned aerial vehicle cooperative task allocation model, the fitness value calculation formula is as follows:
fitnessi＝f(Xi)
Wherein,
Fitness i represents the fitness value of the ith agent;
f (X i) represents a multiple unmanned task allocation objective function for individual X i;
A3. the seagull optimization algorithm is improved by a stagnation judging strategy to judge whether the seagull optimization algorithm falls into local optimization; the stagnation judging strategy judges whether the local optimum is trapped or not by comparing whether the optimal average positions of the front population and the back population are the same or not, and the judging strategy is as follows:
Wherein,
F stag is a flag indicating whether the population falls into a local optimum;
X mean represents the individual optimal average position;
x' mean represents the individual optimal average position of the last iteration number;
X lbest,i represents the current optimal position of the ith seagull;
A4. Updating the gull population position in the gull optimization algorithm; and according to the judging result of the stagnation judging strategy, updating the gull population position by adopting two strategies:
strategy 1, when F stag =0, shows that the seagull population is not trapped into local optimum, and the individual seagull population updates the position information through the migration behavior and the attack behavior of the standard seagull optimization algorithm;
Strategy 2, when F stag =1, representing that the seagull population falls into local optimum, taking the first half population with the minimum fitness value as the dominant population, calculating a probability distribution model through the dominant population of the current iteration times and a weighted maximum likelihood estimation method, and generating a new offspring population according to the probability distribution model;
A5. judging whether an iteration ending condition is met; if the iteration number reaches the maximum iteration number, finishing the iteration, outputting an optimal solution, and decoding the optimal solution into a multi-unmanned aerial vehicle task allocation scheme; otherwise, continuing to A2;
S4, solving a multi-unmanned aerial vehicle collaborative task allocation problem based on an improved seagull optimization algorithm;
s5, analyzing the result of the multi-unmanned aerial vehicle collaborative task allocation.
2. The multi-unmanned aerial vehicle collaborative task allocation method based on the Gaussian distribution seagull optimization algorithm according to claim 1, wherein in S4, the method comprises the following steps:
B1. initializing a multi-unmanned aerial vehicle collaborative task allocation scene;
B2. Establishing a mapping relation between an individual position and a task allocation result through a coding mode based on a real number vector;
B3. determining the number of gull population and the maximum iteration number of algorithm optimization;
B4. Calculating a fitness value function;
B5. updating the gull population position;
B6. And judging whether a termination condition is reached.
3. The multi-unmanned aerial vehicle collaborative task allocation method based on the Gaussian distribution gull optimization algorithm according to claim 2, wherein in B1, the NU unmanned aerial vehicle is assumed to execute an attack task in a known airspace, and NT targets in a task area need to be hit;
In the step B2, the optimal dimension of the multi-unmanned aerial vehicle collaborative task allocation problem is the target number required to be hit, the dimension number of an individual corresponds to the target number, the search space range of the solution is (1, N U +1), the integral part of the individual position corresponds to the number of the UAV, and the decimal part corresponds to the order of executing the task of the same UAV in ascending order;
In the step B3, initializing a seagull population according to a coding mode based on a real number vector, generating a uniformly distributed chaotic sequence in a [0,1] interval by adopting an Arnold chaotic mapping strategy and searching the initialized seagull population by utilizing the chaotic sequence generated by the Arnold chaotic mapping strategy;
in the step B4, decoding the individual position information of the population into a task allocation mapping relation, and converting the multi-constraint multi-unmanned aerial vehicle task allocation problem into an unconstrained optimization problem by adopting a penalty function method:
fitness＝maxJ
the task allocation problem of the multiple unmanned aerial vehicles is a maximized problem, and the larger the objective function value of the model is, the larger the fitness value of the population individuals is, and the better the task allocation solution is;
in B5, whether the sea gull population is trapped in local optimum is judged by comparing whether the optimal average positions of the sea gull population before and after are the same, when F stag =0, the sea gull population is not trapped in local optimum, policy 1 is adopted as a stagnation judgment policy, when F stag =0, the sea gull population is trapped in local optimum, and policy 2 is adopted as a stagnation judgment policy;
In step B6, judging whether the termination condition is met, outputting an optimal task allocation scheme, if the iteration number reaches the maximum iteration number, ending the iteration, outputting an optimal solution, and decoding the optimal solution into a multi-unmanned aerial vehicle task allocation scheme; otherwise, continue B4.
4. The multi-unmanned aerial vehicle collaborative task allocation method based on the Gaussian distribution seagull optimization algorithm according to claim 3, wherein in S5, the method comprises the following steps:
C1. Comparing and simulating different distance models;
C2. task allocation model feasibility simulation;
C3. and (5) expanding the task scale simulation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410079406.7A CN117590757B (en) | 2024-01-19 | 2024-01-19 | Multi-unmanned aerial vehicle cooperative task allocation method based on Gaussian distribution sea-gull optimization algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410079406.7A CN117590757B (en) | 2024-01-19 | 2024-01-19 | Multi-unmanned aerial vehicle cooperative task allocation method based on Gaussian distribution sea-gull optimization algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117590757A CN117590757A (en) | 2024-02-23 |
| CN117590757B true CN117590757B (en) | 2024-05-07 |
Family
ID=89917032
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410079406.7A Active CN117590757B (en) | 2024-01-19 | 2024-01-19 | Multi-unmanned aerial vehicle cooperative task allocation method based on Gaussian distribution sea-gull optimization algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117590757B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2362951A2 (en) * | 2008-06-27 | 2011-09-07 | Raytheon Company | Apparatus and method of controlling an unmanned vehicle |
| CN110018694A (en) * | 2019-05-20 | 2019-07-16 | 成都航空职业技术学院 | It is a kind of for the unmanned plane collaboratively searching method with discontinuous observating characteristic target |
| RU199963U1 (en) * | 2020-03-11 | 2020-09-29 | Акционерное общество "Стелла-К" | Airborne personal protection system of an aircraft from the damaging effects of portable anti-aircraft missile systems |
| WO2021088311A1 (en) * | 2019-11-04 | 2021-05-14 | 中国科学院自动化研究所 | Multi-unmanned aerial vehicle collaborative operation-based automatic inspection method and system for bridges |
| CN113359508A (en) * | 2021-05-25 | 2021-09-07 | 中国人民解放军陆军航空兵学院 | Helicopter/unmanned aerial vehicle weapon cooperative data chain communication system simulation modeling method, medium and system |
| CN115718463A (en) * | 2022-07-21 | 2023-02-28 | 西北工业大学 | Workshop full-element model for airplane assembly |
| CN115808933A (en) * | 2022-11-30 | 2023-03-17 | 南京信息工程大学 | Unmanned aerial vehicle path planning method based on quantum seagull algorithm |
| CN115903514A (en) * | 2022-12-11 | 2023-04-04 | 中国舰船研究设计中心 | Robust scheduling method for complex equipment resource project cluster in random working |
| WO2023101228A1 (en) * | 2021-11-30 | 2023-06-08 | 네이버랩스 주식회사 | Robot-compatible building, and cooperation method and system using plurality of robots |
| CN116991500A (en) * | 2023-08-03 | 2023-11-03 | 南京邮电大学 | Edge/cloud task unloading method based on improved seagull algorithm |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130339969A1 (en) * | 2012-06-19 | 2013-12-19 | Nmetric, Llc | Scheduling and Decision System |
| CA3035970A1 (en) * | 2016-09-09 | 2018-03-15 | Walmart Apollo, Llc | Apparatus and method for unmanned flight |
-
2024
- 2024-01-19 CN CN202410079406.7A patent/CN117590757B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2362951A2 (en) * | 2008-06-27 | 2011-09-07 | Raytheon Company | Apparatus and method of controlling an unmanned vehicle |
| CN110018694A (en) * | 2019-05-20 | 2019-07-16 | 成都航空职业技术学院 | It is a kind of for the unmanned plane collaboratively searching method with discontinuous observating characteristic target |
| WO2021088311A1 (en) * | 2019-11-04 | 2021-05-14 | 中国科学院自动化研究所 | Multi-unmanned aerial vehicle collaborative operation-based automatic inspection method and system for bridges |
| RU199963U1 (en) * | 2020-03-11 | 2020-09-29 | Акционерное общество "Стелла-К" | Airborne personal protection system of an aircraft from the damaging effects of portable anti-aircraft missile systems |
| CN113359508A (en) * | 2021-05-25 | 2021-09-07 | 中国人民解放军陆军航空兵学院 | Helicopter/unmanned aerial vehicle weapon cooperative data chain communication system simulation modeling method, medium and system |
| WO2023101228A1 (en) * | 2021-11-30 | 2023-06-08 | 네이버랩스 주식회사 | Robot-compatible building, and cooperation method and system using plurality of robots |
| CN115718463A (en) * | 2022-07-21 | 2023-02-28 | 西北工业大学 | Workshop full-element model for airplane assembly |
| CN115808933A (en) * | 2022-11-30 | 2023-03-17 | 南京信息工程大学 | Unmanned aerial vehicle path planning method based on quantum seagull algorithm |
| CN115903514A (en) * | 2022-12-11 | 2023-04-04 | 中国舰船研究设计中心 | Robust scheduling method for complex equipment resource project cluster in random working |
| CN116991500A (en) * | 2023-08-03 | 2023-11-03 | 南京邮电大学 | Edge/cloud task unloading method based on improved seagull algorithm |
Non-Patent Citations (2)
| Title |
|---|
| Cooperative Task Assignment Algorithm of Manned/ Unmanned Aerial Vehicle in Uncertain Environment;Li Bo 等;《2017 IEEE 2nd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC)》;20180208;1119-1123 * |
| 多UCAV动态协同任务规划建模与滚动优化方法研究;霍霄华;《中国博士学位论文全文数据库信息科技辑》;20090715(第7期);I138-130 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117590757A (en) | 2024-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111722643B (en) | Unmanned aerial vehicle cluster dynamic task allocation method imitating wolf colony cooperative hunting mechanism | |
| CN113741508B (en) | Unmanned aerial vehicle task allocation method based on improved wolf pack algorithm | |
| CN108549402A (en) | Unmanned aerial vehicle group method for allocating tasks based on quantum crow group hunting mechanism | |
| CN113009934A (en) | Multi-unmanned aerial vehicle task dynamic allocation method based on improved particle swarm optimization | |
| CN104573812A (en) | Uninhabited combat air vehicle route path determining method based on PGSO (Particle-Glowworm Swarm Optimization) algorithm | |
| CN112070418B (en) | Weapon target distribution method of multi-target whale optimization algorithm | |
| CN109269502A (en) | A kind of no-manned plane three-dimensional Route planner based on more stragetic innovation particle swarm algorithms | |
| Li et al. | An adaptive whale optimization algorithm using Gaussian distribution strategies and its application in heterogeneous UCAVs task allocation | |
| CN111553469A (en) | Wireless sensor network data fusion method, device and storage medium | |
| CN111121784B (en) | Unmanned reconnaissance aircraft route planning method | |
| CN111797966B (en) | Multi-machine collaborative global target distribution method based on improved flock algorithm | |
| CN111709511A (en) | Harris eagle optimization algorithm based on random unscented Sigma point variation | |
| Liu et al. | Community detection utilizing a novel multi-swarm fruit fly optimization algorithm with hill-climbing strategy | |
| CN115047907B (en) | Air isomorphic formation command method based on multi-agent PPO algorithm | |
| CN116933948A (en) | Prediction method and system based on improved seagull algorithm and back propagation neural network | |
| CN115494873A (en) | Heterogeneous multi-unmanned aerial vehicle cooperative task allocation method based on Monte Carlo tree search architecture under time sequence constraint | |
| CN114662638A (en) | Mobile robot path planning method based on improved artificial bee colony algorithm | |
| CN110766125A (en) | Multi-target weapon-target allocation method based on artificial fish swarm algorithm | |
| CN115730743A (en) | Battlefield combat trend prediction method based on deep neural network | |
| Jing et al. | Cooperative task assignment for heterogeneous multi-UAVs based on differential evolution algorithm | |
| CN117590757B (en) | Multi-unmanned aerial vehicle cooperative task allocation method based on Gaussian distribution sea-gull optimization algorithm | |
| CN117421969A (en) | Wind speed prediction method for optimizing LSSVM model based on BSSA algorithm | |
| CN116841707A (en) | Sensor resource scheduling method based on HDABC algorithm | |
| CN116088586A (en) | Method for planning on-line tasks in unmanned aerial vehicle combat process | |
| Espinós Longa et al. | Swarm Intelligence in Cooperative Environments: Introducing the N-Step Dynamic Tree Search Algorithm |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |