Improved multi-objective ant colony method based on diffusion mechanism for constructing unmanned aerial vehicle networks

Abstract

The application discloses an improved multi-target ant colony method based on a diffusion mechanism for constructing a UAV network, and comprises the following steps: firstly, a UAV network model is established, and network nodes and network architecture are defined; then, network length and network unreliability are defined as optimization indexes on the basis of the UAV network model; after that, a three-dimensional space diffusion mechanism is optimized and converted into a one-dimensional diffusion mechanism; finally, the diffusion mechanism is applied to a multi-target ant colony algorithm, and a multi-target ant colony algorithm based on the diffusion mechanism is proposed. The application can effectively solve a high-efficiency and reliable UAV network topology structure, solves the problem of solving the UAV network topology structure to a certain extent, and greatly improves the solving efficiency.

Description

Technical Field

[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) technology, specifically relating to an improved multi-target ant colony method based on diffusion mechanisms. Background Technology

[0002] Unmanned aerial vehicle (UAV) networks refer to communication networks composed of one or more UAV nodes, characterized by complex structures, dynamically changing topologies, and diverse nodes. Constructing efficient and reliable UAV networks within a given spatial location is crucial. On one hand, an efficient UAV network ensures rapid information transmission between UAVs; on the other hand, a reliable UAV network can accommodate reliability differences among various UAVs, guaranteeing effective information propagation throughout the network. Existing work largely focuses on using heuristic algorithms for cooperative path planning in UAV swarms, studying the maximum coverage problem of UAV swarms, and researching energy-efficient routing protocols. Therefore, constructing reliable and efficient UAV networks remains an urgent problem to be solved. Summary of the Invention

[0003] To overcome the shortcomings of existing technologies, this invention provides an improved multi-objective ant colony method based on a diffusion mechanism for constructing UAV networks. First, a UAV network model is established, defining network nodes and network architecture. Then, based on the UAV network model, network length and network unreliability are defined as optimization indicators. Next, the three-dimensional diffusion mechanism is optimized and transformed into a one-dimensional diffusion mechanism. Finally, the diffusion mechanism is applied to a multi-objective ant colony algorithm, proposing a diffusion-based multi-objective ant colony algorithm. This invention can effectively solve efficient and reliable UAV network topologies, solving the problem of UAV network topology solving to a certain extent and greatly improving the efficiency of the solution.

[0004] The technical solution adopted by this invention to solve its technical problem includes the following steps:

[0005] Step 1: Establish a drone network model;

[0006] Assuming a drone swarm has n nodes, the entire network can be represented as an undirected graph G = (V, E), where V represents the set of nodes and E represents the set of edges; the location information of node i can be represented by (x... i y i , z i )express;

[0007] Step 2: Define network efficiency and reliability metrics;

[0008] Objective 1: Construct a network topology that minimizes the overall network length and ensures that each node has at least two reachable paths.

[0009]

[0010]

[0011] in, E L Let be the set of edges whose length pheromone concentration is greater than or equal to the average length pheromone concentration of the entire network. The first constraint means that the sum of the lengths of all edges needs to be minimized, and the second constraint means that there must be at least two non-intersecting paths between each node.

[0012] Objective 2: Construct a network topology that maximizes overall network reliability (i.e., minimizes unreliability) and ensures that each node has two reachable paths.

[0013]

[0014]

[0015] Where R(i,j) is the unreliability from node i to node j, R(i,j) = 1 / (i+j), E R Let f(x) be the set of edges whose unreliability pheromone concentration is higher than or equal to the average unreliability pheromone concentration of the entire network.

[0016] Step 3: Diffusion mechanism;

[0017] Assuming that non-directly adjacent edges do not interfere with each other during pheromone diffusion; by setting a pheromone diffusion threshold d, pheromone diffusion is performed on any two adjacent edges, and then the diffusion concentrations of multiple adjacent edges are superimposed and calculated using the finite difference method.

[0018] The formula for calculating the pheromone concentration diffused to edge (i, j) is as follows:

[0019]

[0020] Where D is the diffusion coefficient, τ(i,j) is the pheromone concentration of edge (i,j), k is all nodes directly connected to point j and τ(k,j)>d, m is all nodes directly connected to point i and τ(m,i)>d, Δx is the distance or unreliability from point k to point j, and Δy is the distance or unreliability from point m to point i.

[0021] Step 4: Multi-objective ant colony optimization algorithm based on diffusion mechanism;

[0022] (1) Define the problem: Define the optimization function, network length, and network unreliability;

[0023] (2) Initialize the drone node distance matrix L and the drone node unreliability matrix R, where the value of each element L(i,j) in L represents the actual distance between drone i and drone j, and the value of each element R(i,j) in R represents the unreliability between drone i and drone j;

[0024] (3) Initialize the ant population and pheromone matrix: Set the number of ant populations to 3 and the number of pheromone matrices to 2. The first two ant populations are used to optimize the objective function network length and network unreliability, respectively. The third ant population optimizes both network length and network unreliability. The two pheromone matrices are used to update ant population 1 and ant population 2, respectively.

[0025] (4) Choosing the ant's next path: Calculate the ant's next path based on the pheromone concentration at the current position and the heuristic matrix; the probability of the ant moving to the next node is:

[0026]

[0027] Where τ(i,j) is the pheromone concentration from position i to position j, η(i,j) is the heuristic information from position i to position j, α is the pheromone factor that controls the importance of pheromones, and β is the heuristic factor that controls the importance of heuristic pheromones.

[0028] (5) Update the first ant colony pheromone:

[0029]

[0030] Update the second ant colony pheromone:

[0031]

[0032] Update the third ant colony pheromone:

[0033] τ LR (i, j) = τ L (i, j) + τ R (i, j)

[0034] (6) Disseminate the pheromone of the first ant colony:

[0035]

[0036] Disperse the pheromone of the second ant colony:

[0037]

[0038] (7) Determine the stopping condition: Determine whether the stopping condition of reaching the maximum number of iterations is met;

[0039] (8) Output results: Output the optimal solution of ant population 3 that satisfies the stopping condition as the output of the algorithm, and draw the network topology.

[0040] (9) Return to (4): If the stopping condition is not met, return to (4) to proceed to the next iteration;

[0041] Step 5: Visualize the results;

[0042] Visualize the algorithm iteration results and network topology, and save the relevant result graphs.

[0043] Preferably, step 5 uses Python's Matplotlib library to visualize the algorithm iteration results and network topology, and saves the relevant result graphs.

[0044] The beneficial effects of this invention are as follows:

[0045] 1. This invention is of great significance for constructing efficient and reliable UAV network topologies. By utilizing a multi-objective ant colony algorithm based on a diffusion mechanism to simultaneously optimize network length and network unreliability, it can effectively solve the problem of network topology construction under multi-objective optimization functions.

[0046] 2. Dynamic changes in network topology are a key characteristic of UAV swarm networks, and constructing the network topology within a limited timeframe is both a key research focus and a challenge. The multi-objective ant colony algorithm based on a diffusion mechanism proposed in this invention can rapidly construct UAV networks, significantly optimizing solution efficiency and converging to the optimal solution within a short number of iterations. Attached Figure Description

[0047] Figure 1 This is a flowchart of the multi-target ant colony algorithm based on diffusion mechanism of the present invention.

[0048] Figure 2 The results of the multi-target ant colony algorithm and the multi-target ant colony algorithm based on the diffusion mechanism are shown in the figure. (a) Results of the multi-target ant colony algorithm, (b) Results of the multi-target ant colony algorithm based on the diffusion mechanism. Detailed Implementation

[0049] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0050] To address the problems of existing technologies, this invention proposes a multi-objective ant colony optimization algorithm based on a diffusion mechanism for constructing the topology of unmanned aerial vehicle (UAV) networks. This algorithm aims to achieve the dual goals of efficient communication and improved network reliability in UAV networks. By incorporating heuristic search based on the ant colony algorithm and the diffusion mechanism, it can quickly and effectively optimize the connection relationships between UAV nodes, significantly improving the efficiency of generating network topologies that adapt to task requirements. This algorithm provides superior network communication performance and reliable communication adaptable to the diversity of UAV types, offering an effective method for the rapid deployment and optimization of UAV networks.

[0051] like Figure 1 As shown, a method for constructing UAV network topology based on a multi-target ant colony algorithm with a diffusion mechanism includes the following steps:

[0052] Step 1: Establish a drone network model;

[0053] A drone swarm can be abstracted as a set of points distributed in three-dimensional space. Assuming there are n nodes in the drone swarm, each node can be represented by (x... i y i , z i The node's position in three-dimensional space is represented by its index, and its reliability is indicated by its number. The connections between nodes can be represented as an undirected graph G = (V, E), where V represents the set of nodes and E represents the set of edges.

[0054] Step 2: Define network efficiency and reliability metrics;

[0055] In drone networks, network efficiency can be measured by network length. In an undirected graph G, the length of each edge (i, j) ∈ E can be represented by Euclidean distance:

[0056]

[0057] The first objective of this invention is to construct a network topology that minimizes the overall network length and ensures that each node has two reachable paths:

[0058]

[0059]

[0060] in, E L Let be the set of edges whose length pheromone concentration is greater than or equal to the average length pheromone concentration of the entire network. The first constraint means that the sum of the lengths of all edges needs to be minimized, and the second constraint means that there must be at least two non-intersecting paths between each node.

[0061] Meanwhile, the reliability of node i is defined as i, and the unreliability of the edge from node i to node j is defined as R(i, j) = 1 / (i+j).

[0062] The second objective of this invention is to construct a network topology that maximizes the reliability of the entire network, i.e. minimizes its unreliability, and includes two reachable paths between each node:

[0063]

[0064]

[0065] Among them, E R Let f(x) be the set of edges whose unreliability pheromone concentration is higher than or equal to the average unreliability pheromone concentration of the entire network.

[0066] Step 3: Diffusion mechanism;

[0067] To simplify the complexity and time required to solve the diffusion equation in three-dimensional space, this invention proposes a method that transforms the pheromone diffusion problem in three-dimensional space into a problem that can be solved in one-dimensional space. In this method, it is assumed that non-directly adjacent edges will not interfere with each other during the pheromone diffusion process.

[0068] First, based on the set pheromone diffusion threshold d, adjacent edges within a specified pheromone concentration range are selected for pheromone diffusion. This means that diffusion will only occur when the pheromone concentration of the edges is within a certain range.

[0069] Next, for these adjacent edges, their diffusion concentrations are summed. Specifically, the sum of the pheromone concentrations diffused to that edge from each of these edges is calculated and used as the pheromone diffusion concentration diffused to that edge from the entire network.

[0070] Finally, to solve the pheromone diffusion problem in one-dimensional space, this invention employs the finite difference method. The finite difference method is a commonly used numerical method for approximating partial differential equations. It discretizes the space into several nodes and uses the differences between the nodes to estimate the derivatives. By applying the finite difference method in one-dimensional space, the diffusion concentration at the edges can be effectively obtained, improving computational efficiency and making the pheromone diffusion problem easier to handle and optimize.

[0071] Update the pheromone concentration that diffuses throughout the network to edge (i, j):

[0072]

[0073] Where D is the diffusion coefficient, τ(i,j) is the pheromone concentration of edge (i,j), k is all nodes directly connected to point j and τ(k,j)>d, m is all nodes directly connected to point i and τ(m,i)>d, Δx is the distance (or unreliability) from point k to point j, and Δy is the distance (or unreliability) from point m to point i.

[0074] Step 4: Multi-objective ant colony optimization algorithm based on diffusion mechanism;

[0075] (1) Define the problem: Define the optimization function, network length, and network unreliability.

[0076] (2) Initialize the drone node distance matrix L and the drone node unreliability matrix R, where the value of each element L(i,j) in L represents the actual distance between drone i and drone j, and the value of each element R(i,j) in R represents the unreliability between drone i and drone j.

[0077] (3) Initialize ant colonies and pheromone matrices: Each colony considers a different objective and uses its own pheromone structure and heuristic information to construct a solution; an additional multi-objective ant colony is added to optimize all objectives. The number of ant colonies is set to 3, and the number of pheromone matrices is set to 2. The first two ant colonies are used to optimize the objective function network length and network unreliability, respectively. The third ant colony optimizes both network length and network unreliability simultaneously. The solution of the third colony is only updated when both network length and network unreliability are better than the known optimal solution.

[0078] (4) Choosing the ant's next path: Calculate the ant's next path based on the pheromone concentration at its current position and the heuristic matrix. The probability of the ant moving to the next node is:

[0079]

[0080] Where τ(i,j) is the pheromone concentration from position i to position j, η(i,j) is the heuristic information from position i to position j, α is the pheromone factor, set to 1, which controls the importance of pheromones; β is the heuristic factor, set to 0.5, which controls the importance of heuristic pheromones.

[0081] The selection strategy of ant colony 1 is to make probabilistic selections based on the pheromone matrix of ant colony 1, which is used to optimize the network length. This allows the ant selection strategy to optimize the network length and reach the optimal solution.

[0082] The selection strategy of ant colony 2 is to make probabilistic selections based on the pheromone matrix of ant colony 2, which is used to optimize the network unreliability. This allows the ant selection strategy to optimize the network unreliability and reach the optimal solution.

[0083] The selection strategy for ant colony 3 is to make probabilistic selections based on the fused pheromone matrix of ant colonies 1 and 2, while simultaneously optimizing the network length and network unreliability, so that the network topology it constructs can achieve the optimal solution on both objective functions at the same time.

[0084] (5) After the ants have gone through random walks, update the pheromone concentration of the ant population.

[0085] Update the first ant colony pheromone:

[0086]

[0087] Update the second ant colony pheromone:

[0088]

[0089] Update the third ant colony pheromone:

[0090] τ LR (i, j) = τ L (i, j) + τ R (i, j)

[0091] (6) Disseminate the pheromone of the first ant colony:

[0092]

[0093] Disperse the pheromone of the second ant colony:

[0094]

[0095] The pheromone of the third ant colony is obtained by weighted summation of the first and second ant colonies, and no pheromone diffusion is required.

[0096] (7) Determine the stopping condition: Determine whether the stopping condition of reaching the maximum number of iterations is met.

[0097] (8) Output results: Output the optimal solution of ant population 3 that satisfies the stopping condition as the output of the algorithm, and draw the network topology.

[0098] (9) Return to (4): If the stopping condition is not met, return to (4) to proceed to the next iteration.

[0099] This invention uses ant colony optimization and multi-objective ant colony optimization as references, and designs three sets of comparative experiments: ant colony optimization, diffusion-based ant colony optimization, and multi-objective ant colony optimization. Experiments were conducted on two optimization indicators respectively.

[0100] Step 5: Visualize the results.

[0101] The algorithm iteration results and network topology are visualized using Python's Matplotlib library. The iteration process of the optimization function network length and network unreliability of three ant colonies is plotted, and the network topology generated by ant colony 1, ant colony 2 and ant colony 3 is plotted and the relevant results are saved. Specific implementation examples:

[0103] 1. Establish a drone network model;

[0104] A drone swarm can be abstracted as a set of points distributed in three-dimensional space. Let's set the number of drones in the swarm to 10, and each node can be represented by (x...). i y i , z i The number () represents its position in three-dimensional space, and the node's index indicates its reliability. The connections between nodes can be represented as an undirected graph G = (V, E), where V represents the set of nodes {0, 1, 2, 3, 4, 5, 6, 7, 8, 9}, and E represents the set of edges.

[0105] 2. Define network efficiency and reliability metrics;

[0106] In drone networks, network efficiency can be measured by network length. In an undirected graph G, the length of each edge (i, j) ∈ E can be represented by Euclidean distance:

[0107]

[0108] The first objective of this invention is to construct a network topology that minimizes the overall network length and ensures that each node has two reachable paths:

[0109]

[0110]

[0111] Among them, E L Let be the set of edges whose length pheromone concentration is greater than or equal to the average length pheromone concentration of the entire network. The first constraint means that the sum of the lengths of all edges needs to be minimized, and the second constraint means that there must be at least two non-intersecting paths between each node.

[0112] Meanwhile, the reliability of node i is defined as i, and the unreliability of the edge from node i to node j is defined as R(i, j) = 1 / (i+j).

[0113] The second objective of this invention is to construct a network topology that maximizes the reliability of the entire network, i.e. minimizes its unreliability, and includes two reachable paths between each node:

[0114]

[0115]

[0116] Among them, E R Let f(x) be the set of edges whose unreliability pheromone concentration is higher than or equal to the average unreliability pheromone concentration of the entire network.

[0117] 3. Diffusion mechanism;

[0118] To simplify the complexity and time required to solve the diffusion equation in three-dimensional space, this invention proposes a method that transforms the pheromone diffusion problem in three-dimensional space into a problem that can be solved in one-dimensional space. In this method, it is assumed that non-directly adjacent edges will not interfere with each other during the pheromone diffusion process.

[0119] First, based on the set pheromone diffusion threshold d, which is the average pheromone concentration in the pheromone matrix, adjacent edges within a specified pheromone concentration range are selected for pheromone diffusion. This means that diffusion will only occur when the pheromone concentration of an edge is within a certain range.

[0120] Next, for these adjacent edges, their diffusion concentrations are summed. Specifically, the sum of the pheromone concentrations diffused to that edge from each of these edges is calculated and used as the pheromone diffusion concentration diffused to that edge from the entire network.

[0121] Finally, to solve the pheromone diffusion problem in one-dimensional space, this invention employs the finite difference method. The finite difference method is a commonly used numerical method for approximating partial differential equations. It discretizes the space into several nodes and uses the differences between the nodes to estimate the derivatives. By applying the finite difference method in one-dimensional space, the diffusion concentration at the edges can be effectively obtained, improving computational efficiency and making the pheromone diffusion problem easier to handle and optimize.

[0122] Update the pheromone concentration that diffuses throughout the network to edge (i, j):

[0123]

[0124] Where D is the diffusion coefficient, τ(i,j) is the pheromone concentration of edge (i,j), k is all nodes directly connected to point j and τ(k,j)>d, m is all nodes directly connected to point i and τ(m,i)>d, Δx is the distance (or unreliability) from point k to point j, and Δy is the distance (or unreliability) from point m to point i.

[0125] 4. Multi-objective ant colony optimization algorithm based on diffusion mechanism;

[0126] (1) Define the problem: Define the optimization function, network length, and network unreliability.

[0127] (2) Initialize the drone node distance matrix L and the drone node unreliability matrix R, where the value of each element L(i,j) in L represents the actual distance between drone i and drone j, and the value of each element R(i,j) in R represents the unreliability between drone i and drone j.

[0128] (3) Initialize ant colonies and pheromone matrices: Each colony considers a different objective and uses its own pheromone structure and heuristic information to construct a solution; an additional multi-objective ant colony is added to optimize all objectives. The number of ant colonies is set to 3, and the number of pheromone matrices is set to 2. The first two ant colonies are used to optimize the objective function network length and network unreliability, respectively. The third ant colony optimizes both network length and network unreliability simultaneously. The solution of the third colony is only updated when both network length and network unreliability are better than the known optimal solution.

[0129] (4) Choosing the ant's next path: Calculate the ant's next path based on the pheromone concentration at its current position and the heuristic matrix. The probability of the ant moving to the next node is:

[0130]

[0131] Where τ(i,j) is the pheromone concentration from position i to position j, η(i,j) is the heuristic information from position i to position j, α is the pheromone factor, set to 1, which controls the importance of pheromones; β is the heuristic factor, set to 0.5, which controls the importance of heuristic pheromones.

[0132] The selection strategy of ant colony 1 is to make probabilistic selections based on the pheromone matrix of ant colony 1, which is used to optimize the network length. This allows the ant selection strategy to optimize the network length and reach the optimal solution.

[0133] The selection strategy of ant colony 2 is to make probabilistic selections based on the pheromone matrix of ant colony 2, which is used to optimize the network unreliability. This allows the ant selection strategy to optimize the network unreliability and reach the optimal solution.

[0134] The selection strategy for ant colony 3 is to make probabilistic selections based on the fused pheromone matrix of ant colonies 1 and 2, while simultaneously optimizing the network length and network unreliability, so that the network topology it constructs can achieve the optimal solution on both objective functions at the same time.

[0135] (5) After the ants have gone through random walks, update the pheromone concentration of the ant population.

[0136] Update the first ant colony pheromone:

[0137]

[0138] Update the second ant colony pheromone:

[0139]

[0140] Update the third ant colony pheromone:

[0141] τ LR (i, j) = τ L (i, j) + τ R (i, j)

[0142] (6) Disseminate the pheromone of the first ant colony:

[0143]

[0144] Disperse the pheromone of the second ant colony:

[0145]

[0146] The pheromone of the third ant colony is obtained by weighted summation of the first and second ant colonies, and no pheromone diffusion is required.

[0147] (7) Determine the stopping condition: Determine whether the stopping condition of reaching the maximum number of iterations is met.

[0148] (8) Output results: Output the optimal solution of ant population 3 that satisfies the stopping condition as the output of the algorithm, and draw the network topology.

[0149] (9) Return to (4): If the stopping condition is not met, return to (4) to proceed to the next iteration.

[0150] 5. Testing Phase

[0151] This invention was tested on UAV networks of sizes 10 and 20 to verify the effectiveness of the multi-target ant colony algorithm based on a diffusion mechanism. This invention can rapidly construct efficient and reliable topologies on UAV networks of different sizes.

[0152] like Figure 2 The results are shown in the comparison charts of the multi-target ant colony algorithm and the multi-target ant colony algorithm based on the diffusion mechanism, where (a) is the result of the multi-target ant colony algorithm and (b) is the result of the multi-target ant colony algorithm based on the diffusion mechanism.

[0153] This invention proposes a diffusion-based multi-objective ant colony algorithm for rapidly constructing efficient and reliable UAV network topologies. This invention applies the ant colony algorithm to the task of constructing UAV network topologies and proposes a multi-objective ant colony algorithm to simultaneously optimize network efficiency and reliability. Considering the dynamic characteristics of UAV swarms, this invention proposes a pheromone diffusion mechanism based on diffusion equations and applies it to the multi-objective ant colony algorithm, resulting in a diffusion-based multi-objective ant colony algorithm. This method can significantly improve the efficiency of constructing UAV network topologies while optimizing multiple objective functions. The diffusion-based multi-objective ant colony algorithm proposed in this invention is not only suitable for the rapid construction of efficient and reliable UAV networks, but can also be widely applied to other tasks. The number of objective functions to be optimized can be arbitrarily specified, and the diffusion mechanism is plug-and-play.

Claims

1. A diffusion-based improved multi-target ant colony method for constructing unmanned aerial vehicle (UAV) networks, characterized in that, Includes the following steps: Step 1: Establish a drone network model; Assuming there are drones in the swarm n Each node represents the entire network as an undirected graph. ,in Represents a set of nodes. Represents the set of edges; nodes i Location information can be used express; Step 2: Define network efficiency and reliability metrics; Objective 1: Construct a network topology with the shortest possible length and two reachable paths between each node. in, , Let be the set of edges whose length pheromone concentration is greater than or equal to the average length pheromone concentration of the entire network. The first constraint means that the sum of the lengths of all edges needs to be minimized, and the second constraint means that there must be at least two non-intersecting paths between each node. Objective 2: Construct a network topology that maximizes overall network reliability (i.e., minimizes unreliability) and ensures that there are two reachable paths between each node. in, For nodes i To the node j The unreliability, , Let f(x) be the set of edges whose unreliability pheromone concentration is higher than or equal to the average unreliability pheromone concentration of the entire network. Step 3: Diffusion mechanism; Assuming that non-directly adjacent edges do not interfere with each other during pheromone diffusion; by setting a pheromone diffusion threshold... d Pheromones are diffused between any two adjacent edges, and the diffusion concentrations of multiple adjacent edges are superimposed and calculated using the finite difference method. Spread to the edge The formula for calculating pheromone concentration is as follows: Where D is the diffusion coefficient, For the edge pheromone concentration, k For all points j Directly connected and The node, m For all points i Directly connected and The node, For point k Time j Distance or unreliability, For point m Time i Distance or unreliability; Step 4: Multi-objective ant colony optimization algorithm based on diffusion mechanism; (1) Define the problem: Define the optimization function, network length, and network unreliability; (2) Initialize the UAV node distance matrix L Unreliability matrix of drone nodes R ,in L Each element The value represents the drone i To drones j The actual distance between them R Each element The value represents the drone i To drones j The unreliability between them; (3) Initialize ant population and pheromone matrix: Set the number of ant populations to 3 and the number of pheromone matrices to 2. The first two ant populations are used to optimize the objective function network length and network unreliability, respectively. The third ant population optimizes both network length and network unreliability. The two pheromone matrices are used to update ant population 1 and ant population 2, respectively. (4) Choosing the ant's next path: Calculate the ant's next path based on the pheromone concentration at the current position and the heuristic matrix; the probability of the ant moving to the next node is: in, From position i Arrive at the location j pheromone concentration, From position i Arrive at the location j Heuristic information, As a pheromone factor, it controls the importance of pheromones; As a heuristic factor, the importance of controlling heuristic pheromones; (5) Update the first ant colony pheromone: Update the second ant colony pheromone: Update the third ant colony pheromone: (6) Disperse the pheromone of the first ant colony: Disperse the pheromone of the second ant colony: (7) Determine the stopping condition: Determine whether the stopping condition of reaching the maximum number of iterations is met; (8) Output results: Output the optimal solution of ant population 3 that satisfies the stopping condition as the output of the algorithm, and draw the network topology; (9) Return to (4): If the stopping condition is not met, return to (4) to proceed to the next iteration; Step 5: Visualize the results; Visualize the algorithm iteration results and network topology, and save the relevant result graphs.

2. The improved multi-target ant colony method based on diffusion mechanism for constructing unmanned aerial vehicle (UAV) networks according to claim 1, characterized in that, Step 5 uses Python's Matplotlib library to visualize the algorithm iteration results and network topology, and saves the relevant result graphs.

Images