A multi-unmanned aerial vehicle network coverage optimization method
By optimizing multi-UAV network coverage using the Levy flight mode and a nonlinear decreasing variable swarm particle number strategy, the problems of slow convergence speed and low coverage are solved, and more efficient network coverage optimization is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
- Filing Date
- 2022-10-25
- Publication Date
- 2026-05-05
AI Technical Summary
Existing multi-UAV network coverage optimization methods suffer from slow convergence speed and low coverage, making it difficult to achieve broad network coverage.
A local population is generated using the Levy flight method, and the original candidate solutions in the equilibrium pool are updated based on the fitness value. Combined with a strategy of nonlinearly decreasing the number of particles in the population, the number of particles in the iteration process is dynamically adjusted to optimize the multi-UAV network coverage problem.
It accelerates the convergence speed of the algorithm, improves the coverage, reduces the computational load, avoids getting trapped in local optima, and achieves more efficient network coverage optimization.
Smart Images
Figure CN115935560B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of multi-UAV network coverage optimization, and specifically to the design and application of a multi-UAV network coverage optimization method. Background Technology
[0002] Using drones to mount communication equipment and establish high-altitude base stations for data relay has become a research hotspot in the engineering field. The coverage area of a single drone is limited; therefore, networking multiple drones can achieve a wider coverage area. Traditional methods, such as virtual force field methods and geometric calculation methods, suffer from low coverage and high computational load when solving wireless network node deployment problems, while intelligent algorithms perform well in this area. However, some existing intelligent algorithms also suffer from slow convergence speed and low coverage when solving network coverage optimization problems, often failing to achieve the expected goals.
[0003] Therefore, it is particularly important to design and verify a practical method for optimizing multi-UAV network coverage. Summary of the Invention
[0004] Purpose of the Invention: This invention proposes a high-performance and easy-to-implement optimization method and verification method for multi-UAV network coverage deployment tasks.
[0005] Technical solution:
[0006] This application provides a method for optimizing network coverage for multiple unmanned aerial vehicles (UAVs), the method comprising the following steps:
[0007] To model the multi-UAV network coverage optimization problem, the position coordinates of the UAVs in a two-dimensional plane are encoded as particles in the optimization algorithm, with the dimension of a single particle set to... dimension, The first row represents the number of drones, and the second row represents the x-coordinate of each drone in the two-dimensional plane. The boundary of the target region is set as the upper and lower bounds of the solution space. The network coverage of the target region is set as the objective function of the optimization algorithm.
[0008] Based on the equilibrium optimization algorithm, a local population is generated for candidate solutions using the Lévy flight method, and the original candidate solutions in the equilibrium pool are updated and replaced according to the fitness value; a nonlinear decreasing variable population particle number strategy is adopted to dynamically adjust the particle number during the iteration process; and the solution optimization is performed for the multi-UAV network coverage problem.
[0009] Decoding the historical best particles obtained by the algorithm yields a schematic diagram of the spatial distribution and network coverage area of the drone swarm.
[0010] Furthermore, the specific process for solving and optimizing the multi-UAV network coverage problem is as follows:
[0011] S1 sets the initial and final population particle counts and initializes the parameters of the improved balance optimization algorithm.
[0012] S2 sets the upper and lower bounds of the solution space based on the network coverage area information, generates the initial particle distribution based on the number of drones, and initializes the equilibrium pool.
[0013] S3 calculates the fitness value of a single particle in the population based on the evaluation function, compares it with the fitness pool, and updates the fitness pool.
[0014] S4 Calculates the exponential term coefficient F and the generation rate G, and updates them according to the algorithm update formula;
[0015] S5 uses the Lévy flight method to generate a local population for candidate solutions and updates and replaces the original candidate solutions in the balance pool based on the fitness value;
[0016] S6 calculates the particle fitness value, adopts a non-linear decreasing strategy for the number of mutant particles, eliminates a set number of particles according to the fitness value, and calculates the number of particles in the next iteration.
[0017] S7 determines whether the maximum number of iterations has been reached. If it has, the optimal solution in the balance pool is output and decoded as the spatial distribution of the UAV in the target area; otherwise, return to step S3.
[0018] Furthermore, the four historically best particles and their average value are selected to form an equilibrium pool:
[0019] (2)
[0020] in, , , , These are the four best particles in history. The average state of the four particles;
[0021] The candidate solutions in the equilibrium pool are used to generate a local population using the Lévy flyby method, as follows:
[0022] (10)
[0023] In the formula, These correspond to the four historically best particles in the equilibrium pool. This represents the average state of the four particles in the original equilibrium pool. , The desired number of particles in the local population. The step length obtained using the Levi flight method:
[0024] (11)
[0025] (12)
[0026] As the attenuation factor, It is a constant. , obey and Gaussian distribution;
[0027] The update rule for the new candidate solution is to generate local populations of the four historically best particles by equation (10), calculate the fitness value of the particles in the local population, compare it with the fitness value of the corresponding candidate solution, and replace the corresponding candidate solution with the particle with the better fitness value.
[0028] Furthermore, during the iteration process, a strategy of non-linearly decreasing the number of particles in the population is adopted, and the number of particles in the population dynamically changes in each iteration as follows:
[0029] (13)
[0030] In the formula, The number of particles in the current iteration. The initial number of particles. The final number of particles in the population. This represents the current iteration number. The maximum number of iterations, This is a rounding function; after each iteration, the number of particles in the population for the next iteration is calculated, and the particles are sorted according to their fitness values. Particles with low fitness values are eliminated, and the remaining particles are used as the initial population for the next iteration.
[0031] Furthermore, the ratio of the union of all drone coverage areas to the target area is defined as the network coverage rate. Using network coverage rate as the evaluation metric, the evaluation function is designed as follows:
[0032] (17)
[0033] in, The coverage area of a single drone. The area represents the target region; when the number of drones is fixed, the higher the coverage rate, the larger the evaluation function value.
[0034] As a preferred embodiment of this application, the following four algorithms are selected and compared with the optimization algorithm, and the performance of the results is analyzed:
[0035] The Equilibrium Optimizer (EO), Cuckoo Search (CS), Particle Swarm Optimization (PSO), and Grey Wolf Optimizer (GWO) algorithms were analyzed with the same number of iterations for the same number of UAVs. The results showed the advantages and disadvantages of the optimization methods.
[0036] Beneficial effects
[0037] The present invention has the following advantages: (1) The candidate solutions in the equilibrium pool are generated by the Lévy flight method to generate a local population. The particles with higher fitness values in the local population are used to replace the original candidate solutions in the equilibrium pool, which takes into account both the global search performance and the local search performance of the algorithm, accelerates the convergence speed of the algorithm, and avoids getting trapped in local optima; (2) In the process of algorithm iteration, a strategy of nonlinearly decreasing the number of particles in the population is introduced to dynamically adjust the number of particles in the population and reduce the computational load of the algorithm; (3) In the Matlab environment, the same experimental parameters are set and compared with other algorithms. The coverage change curve and the final coverage are analyzed to verify the feasibility and effectiveness of the proposed optimization algorithm. Attached Figure Description
[0038] Figure 1 Overall framework diagram for optimizing network coverage for multiple drones;
[0039] Figure 2 Flowchart of the balanced optimization algorithm;
[0040] Figure 3 This is a schematic diagram illustrating the network coverage of a single drone.
[0041] Figure 4 This is a schematic diagram of a regular hexagonal covering.
[0042] Figure 5 This is a flowchart of the algorithm of the present invention;
[0043] Figure 6 A map showing the random distribution of drones;
[0044] Figure 7 Distribution map of drones optimized by the EO algorithm;
[0045] Figure 8 Distribution map of drones optimized using the GWO algorithm;
[0046] Figure 9 Distribution map of drones after optimization of the PSO algorithm;
[0047] Figure 10Distribution map of drones after optimization of the CS algorithm;
[0048] Figure 11 This is a distribution map of drones optimized by the algorithm of this invention;
[0049] Figure 12 This is a graph showing the change in coverage rate.
[0050] Figure 13 This is a distribution map of heterogeneous UAVs after the algorithm of this invention is optimized. Detailed Implementation
[0051] The present invention will now be further described with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.
[0052] This embodiment describes a multi-UAV network coverage optimization method. Following the explanation of the optimization method, a verification method is also provided. Based on given target area information, the number of UAVs, and the UAV network coverage radius, a spatial distribution map of the multiple UAVs and a network coverage map can be obtained, facilitating the rational deployment of UAV formation flights. Building upon the balance optimization algorithm, a local population is generated for candidate solutions using the Lévy flight method, and the original candidate solutions in the balance pool are updated and replaced according to their fitness values. During the algorithm iteration process, a strategy of non-linearly decreasing the number of particles in the population is adopted to eliminate particles with low fitness values, dynamically adjusting the number of particles in the population to reduce the computational load of the algorithm. In a Matlab simulation environment, with a target area and the number of UAVs set, and using the same number of iterations, the algorithm is compared with other optimization algorithms. The coverage change curve and the final coverage are analyzed to verify the performance of the algorithm designed in this patent.
[0053] The specific process for solving and optimizing the multi-UAV network coverage problem is as follows:
[0054] S1 sets the initial and final population particle counts and initializes the parameters of the improved equilibrium optimization algorithm.
[0055] S2 sets the upper and lower bounds of the solution space based on the network coverage area information, generates the initial particle distribution, and initializes the equilibrium pool;
[0056] S3 calculates the fitness value of a single particle in the population based on the evaluation function, compares it with the fitness pool, and updates the fitness pool.
[0057] S4 Calculates the exponential term coefficient F and the generation rate G, and updates them according to the algorithm update formula;
[0058] S5 uses the Lévy flight method to generate a local population for candidate solutions and updates and replaces the original candidate solutions in the balance pool based on the fitness value;
[0059] S6 calculates the fitness value of particles in the population, adopts a strategy of non-linearly decreasing the number of particles in the population, eliminates particles with low fitness values, and calculates the number of particles in the population for the next iteration.
[0060] S7 determines whether the maximum number of iterations has been reached. If so, the optimal solution in the balance pool is output and decoded as the spatial distribution of the UAV within the target area; otherwise, return to step S3.
[0061] The specific implementation plan is as follows:
[0062] Balanced optimization algorithm
[0063] The flowchart of the basic balance optimization algorithm is as follows: Figure 2 As shown, the entire process includes five parts: population initialization, balancing pool selection, exponent calculation, generation rate calculation, and individual update rules.
[0064] Population initialization:
[0065] (1)
[0066] In the formula, For particles in the population, and To solve for the lower and upper bounds of the space, An array with the same dimensions as the particles.
[0067] Selection of equilibration pool:
[0068] (2)
[0069] (3)
[0070] The four best particles in history and the average state of these four particles It forms a balance pool.
[0071] Exponential coefficient calculate:
[0072] (4)
[0073] (5)
[0074] , These are the weighting coefficients. , This is a vector of random numbers, with dimensions equal to the particle dimensions, where the element values are... A random number between [a certain number of points]. This represents the current iteration number. This represents the maximum number of iterations.
[0075] Generation rate :
[0076] (6)
[0077] (7)
[0078] (8)
[0079] The initial value for the generation rate, This is a candidate solution in the equilibrium pool. , for Random numbers between It is a constant.
[0080] Individual update rules:
[0081] (9)
[0082] For the current individual, The candidate solutions selected in equation (6) are given.
[0083] Candidate solution update rules
[0084] The balance optimization algorithm updates the entire population and the candidate solutions in the balance pool through equation (9). In the early stage, the algorithm focuses on global search, while in the later search process it tends to focus on local search. This approach leads to slow convergence speed and easy trapping in local optima.
[0085] The Lévy flight algorithm consists of multiple short-step and a few long-step algorithms. This search method combines local and global search, effectively improving the convergence speed and accuracy of the algorithm and preventing it from getting trapped in local optima. This invention uses the Lévy flight algorithm to update candidate solutions. Each candidate solution in the equilibrium pool is iteratively updated a certain number of times according to the Lévy flight algorithm, generating a new population containing M particles. The candidate solution update method is as follows:
[0086] (10)
[0087] In the formula, These correspond to the four historically best particles in the equilibrium pool. This represents the average state of the four particles in the original equilibrium pool. , The desired number of particles in the local population. The step size obtained by Levi's flight method:
[0088] (11)
[0089] (12)
[0090] As the attenuation factor, It is a constant, typically taking the value of 1.5. , obey and Gaussian distribution, This is the gamma function.
[0091] The update rule for the new candidate solution is to generate local populations of the four historically best particles by equation (10), calculate the fitness value of the particles in the local population, compare it with the fitness value of the corresponding candidate solution, and replace the corresponding candidate solution with the particle with the better fitness value.
[0092] Population size dynamic adjustment strategy
[0093] Increasing the population size improves the algorithm's global search capability and convergence accuracy, but excessively large populations slow down the convergence speed. Decreasing the population size speeds up convergence but makes the algorithm more susceptible to local optima. This paper employs a nonlinear decreasing population particle size strategy. In the early stages of the algorithm, a large population size is maintained to enhance global search capability and avoid local optima. As the number of iterations increases, the population size is gradually reduced to improve local search capability and reduce computational complexity. The population size in each iteration is calculated using the following formula:
[0094] (13)
[0095] In the formula, The number of particles in the current iteration. The initial number of particles. The final number of particles in the population. This represents the current iteration number. The maximum number of iterations, This is the floor function.
[0096] Modeling and Solving the Problem of Multi-UAV Network Coverage
[0097] The coverage area of the drone network is represented using a binary perception model:
[0098] (14)
[0099] in, The coverage probability of the target area. The distance from the target area to the drone. This represents the coverage radius of the drone network. If a given area is within the circle, it is considered that the area can sense the network; otherwise, there is no network signal in that area. Figure 3 This is a diagram illustrating the network coverage of a single drone. The flight altitude of the drone. This refers to the coverage radius of a single drone.
[0100] When the area of the mission area and the area that a single drone can cover are known, the required number of drones can be initially estimated. A hexagonal coverage pattern can be used, such as... Figure 4 As shown, the effective coverage area is relatively large. Theoretically, the effective coverage area of a single drone is:
[0101] (15)
[0102] Therefore, the area of the target region is known. In theory, the number of drones required to achieve full coverage is [number missing]. for:
[0103] (16)
[0104] When multiple drones are networked to provide network services to a target area, assuming all drones fly at the same altitude and are controlled by a ground station, the coverage radius of each drone is equal. The network coverage rate of the target area is defined as the ratio of the union of the areas covered by all drones to the total area of the target area. Therefore, the network coverage rate is related to the spatial distribution of the drones, transforming the optimization objective into solving for the spatial distribution of the drone swarm within the target area. The evaluation function is designed as follows:
[0105] (17)
[0106] in, This refers to the coverage area of a single drone over a target region. Let be the area of the target region. With a fixed number of drones, the higher the coverage rate, the larger the evaluation function value.
[0107] The dimension of a single particle is set to dimension, The first row represents the number of drones, and the second row represents the x-coordinate of each drone in a two-dimensional plane. The boundary of the target region is set as the upper and lower bounds of the solution space. The above optimization algorithm is used to iteratively update the particles, calculate the fitness value, and select the historically best particle for decoding to obtain the spatial distribution map of the drones.
[0108] The steps for applying the improved balance optimization algorithm to the multi-UAV network coverage optimization problem are as follows:
[0109] ① Set parameters such as the initial and final population particle count, and the maximum number of iterations;
[0110] ② Set the upper and lower bounds of the solution space based on the target area information, and generate the initial particle distribution based on the number of UAVs;
[0111] ③ Compare the fitness value of each particle with the solution in the previous balancing pool, select the particle with the larger fitness value, and update the candidate solutions in the balancing pool;
[0112] ④ Calculate the coefficient F of the exponential term and the generation rate G, and iteratively update according to the solution update formula;
[0113] ⑤ The candidate solutions in the equilibrium pool are calculated using the Lévy flight method. Particles with high fitness values in the local population are selected to replace the candidate solutions in the equilibrium pool, and the equilibrium pool is updated.
[0114] ⑥ Calculate the fitness value of particles in the population, eliminate particles with low fitness values, and update the number of particles N in the next iteration;
[0115] ⑦ Determine if the maximum number of iterations has been reached. If it has, output the optimal solution in the balance pool; otherwise, return to step ③. The flowchart of the improved algorithm is as follows: Figure 5 As shown.
[0116] Verification method
[0117] This invention was tested in a Matlab simulation environment. The Balanced Optimization Algorithm (EO), the Cuckoo Algorithm (CS), the Particle Swarm Optimization (PSO), and the Grey Wolf Algorithm (GWO) were selected. Under the same target area, the same number of drones, iterations, and population size were set to obtain coverage curve changes and the final network coverage. A faster change in the coverage curve with the number of iterations and a higher final network coverage indicate better global search performance and faster convergence. Conversely, if the coverage curve stagnates frequently and for long periods during iterations, it indicates that the algorithm is prone to getting trapped in local optima and has poor global search performance. Considering both the coverage curve changes and the final network coverage, the performance of the algorithm in the multi-drone network coverage optimization problem was analyzed.
[0118] Numerical simulation verification
[0119] To verify the feasibility and effectiveness of the multi-UAV network coverage optimization method proposed in this invention, simulation experiments were conducted in the Matlab simulation environment. The target area was set as a 50×50km square region, the network coverage radius of a single UAV was 5km, and the number of UAVs was selected as 35. The initial population size was 70 particles, and the final population size was 30 particles. The parameters of the improved equilibrium optimization algorithm were set as follows: , , , The number of flights for Levi was M=10, and the number of iterations was set to 300. The population size for other optimization algorithms was set to 70, and the number of iterations, target areas, and number of drones were the same as above.
[0120] Figure 6 This is a map showing the random distribution of drones. Asterisks indicate drone locations, and the areas within circles represent the drone network coverage. The map contains several empty areas, and some drones are too close together. Figure 7 This is a spatial distribution map of drones optimized by the EO algorithm. The area of uncovered regions in the map has been greatly reduced, but some areas are still uncovered, and drones are relatively concentrated. Figure 8 The image shows the spatial distribution of drones after optimization using the GWO algorithm. In the image, there are instances where multiple drones are too close together, as well as several empty areas. Figure 9 This is a spatial distribution map of UAVs optimized by the PSO algorithm, and... Figure 7 Similarly, the drone was too close and the upper right corner area was not covered. Figure 10 The image shows the spatial distribution of the UAV after optimization using the CS algorithm. Hollow areas exist in multiple corners of the target area. Figure 11 This is a deployment map of drones optimized by the algorithm proposed in this invention. The drones are evenly distributed in the target area, with only a small portion of the area uncovered.
[0121] Figure 12The graph compares the iteration count and coverage rate of each algorithm. As can be seen from the graph, the algorithm proposed in this invention achieves a final coverage rate of 95.68%, which is 5.68%, 9.72%, 4.88%, and 6.80% higher than the EO, CS, GWO, and PSO algorithms, respectively, significantly higher than other algorithms. In terms of convergence speed, the improved balanced optimization algorithm achieves a coverage rate of 95% after 122 iterations, and then slightly increases due to the local search strategy. While the EO algorithm achieves a higher final coverage rate than the PSO and CS algorithms, its coverage rate curve stagnates multiple times during the iteration process, indicating it repeatedly falls into local optima. The CS algorithm only uses the Levy flight method to update the bird's nest position, resulting in poor global search performance and the lowest final coverage rate among the algorithms. It also exhibits poor local search performance and is prone to falling into local optima. The GWO algorithm has an excessively long global search time in the early stages, causing the coverage rate curve to stagnate multiple times for extended periods. However, its global search performance is good, achieving a final coverage rate of 90.8%, second only to the improved algorithm proposed in this invention. The PSO algorithm has a faster convergence speed than other algorithms, but its global search ability is poor and its convergence accuracy is low.
[0122] In the same square target area mentioned above, the coverage radius of the Class I UAV network is set to 5km and the coverage radius of the Class II UAV network is set to 10km. According to formula (16), the number of Class I UAVs is 15, the number of Class II UAVs is 5, and the total number of UAVs is 20. Other initial parameters remain unchanged. Figure 13 The network coverage diagram calculated under the heterogeneous drone state shows that the algorithm can still cover the target area under the heterogeneous drone state, with a coverage rate of 95.9%, and also effectively reduces the number of drones.
[0123] In summary, the multi-UAV network coverage optimization method proposed in this invention has significantly improved convergence speed and convergence accuracy compared with traditional optimization methods, verifying the feasibility and effectiveness of the algorithm in optimizing network coverage for various configurations of UAVs.
Claims
1. A method for optimizing network coverage for multiple unmanned aerial vehicles (UAVs), characterized in that: The method includes the following steps: To model the multi-UAV network coverage optimization problem, the position coordinates of the UAVs in a two-dimensional plane are encoded as particles in the optimization algorithm, with the dimension of a single particle set to... dimension, The first row represents the number of drones, and the second row represents the x-coordinate of each drone in the two-dimensional plane. The boundary of the target region is set as the upper and lower bounds of the solution space. The network coverage of the target region is set as the objective function of the optimization algorithm. Based on the equilibrium optimization algorithm, a local population is generated for candidate solutions using the Lévy flight method, and the original candidate solutions in the equilibrium pool are updated and replaced according to the fitness value; a nonlinear decreasing strategy for the number of particles in the population is adopted to dynamically adjust the number of particles during the iteration process; and the solution speed of the multi-UAV network coverage problem is optimized. The four best historical particles and their average value are selected to form the equilibrium pool: (2) in, , , , These are the four best particles in history. The average state of the four particles; The candidate solutions in the equilibrium pool are used to generate a local population using the Lévy flyby method, as follows: (10) In the formula, These correspond to the four historically best particles in the equilibrium pool. This represents the average state of the four particles in the original equilibrium pool. , For the desired local population size, The step length obtained using the Levi flight method: (11) (12) As the attenuation factor, It is a constant. , obey and Gaussian distribution, It is the gamma function; The update rule for the new candidate solution is to generate the local populations of the four historically best particles by equation (10), calculate the fitness value of the particles in the local population, compare it with the fitness value of the corresponding candidate solution, and replace the corresponding candidate solution with the particle with the better fitness value. During the iteration process, a non-linearly decreasing strategy for the number of particles in the population is adopted, and the number of particles in the population in each iteration is: (13) In the formula, The number of particles in the current iteration. The initial number of particles. The final number of particles in the population. This represents the current iteration number. The maximum number of iterations, This is the floor function; after each iteration, the number of particles in the population for the next iteration is calculated, and the particles are sorted according to their fitness values. Particles with low fitness values are eliminated, and the remaining particles are used as the initial population for the next iteration. Decoding the historical best particles obtained by the algorithm yields a schematic diagram of the spatial distribution and network coverage area of the drone swarm.
2. The multi-UAV network coverage optimization method according to claim 1, characterized in that: The specific process for solving and optimizing the multi-UAV network coverage problem is as follows: S1 sets the initial and final population particle counts and initializes the parameters of the improved equilibrium optimization algorithm. S2 sets the upper and lower bounds of the solution space and the number of drones based on the network coverage area information, generates the initial population particle distribution, and initializes the balance pool; S3 calculates the fitness value of a single particle in the population based on the evaluation function, compares it with the particles in the balance pool, and updates the balance pool. S4 Calculates the exponential term coefficient F and the generation rate G, and updates them according to the algorithm update formula; S5 generates local population particles for candidate solutions using the Lévy flight method, and updates the candidate solutions in the equilibrium pool based on the fitness value; S6 calculates the particle fitness value, adopts a non-linear decreasing strategy for the number of mutant particles, eliminates a set number of particles according to the fitness value, and calculates the number of particles in the next iteration. S7 determines whether the maximum number of iterations has been reached. If it has, the optimal solution in the balance pool is output and established as the spatial distribution of the UAV within the target area; otherwise, return to step S3.
3. The multi-UAV network coverage optimization method according to claim 2, characterized in that: The ratio of the union of all drone coverage areas to the target area is defined as the network coverage rate. Using the network coverage rate as an evaluation metric, the evaluation function is designed as follows: (17) in, The coverage area of a single drone. The area represents the target region; when the number of drones is fixed, the higher the coverage rate, the larger the evaluation function value.
4. A multi-UAV network coverage optimization method according to claim 2 or 3, characterized in that: The following four algorithms were selected and compared with the multi-UAV network coverage optimization method, and the performance of the results was analyzed: The Equilibrium Optimizer (EO), Cuckoo Search (CS), Particle Swarm Optimization (PSO), and Grey Wolf Optimizer (GWO) algorithms were analyzed with the same number of iterations for the same number of drones. The results showed the advantages and disadvantages of the optimization methods.