Path planning method based on improved ant colony algorithm

Abstract

The invention provides a path planning method based on an improved ant colony algorithm. The method is mainly used for optimizing and improving the efficiency and accuracy of the ant colony algorithm in path planning. A traditional ant colony algorithm has the problems of low convergence speed, easy falling into local optimum and the like. In order to solve the problems, the invention provides a series of improvement measures. First, the present invention initializes pheromone concentrations using non-uniform initialization. Secondly, an enhanced direction guiding mechanism is used, and the convergence speed of the algorithm is increased; in addition, the distance between the to-be-selected node and the target point is added into the heuristic function, so that the purposiveness and efficiency of path searching are improved. Finally, according to the invention, a strategy of combining a plurality of transfer strategies is adopted for transfer, and finally, an optimal path with a shorter distance and a relatively smooth path can be obtained more quickly.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to a path planning method based on an improved ant colony algorithm. Background Art

[0002] Path planning has always been a basic and important problem in many fields such as intelligent transportation systems, robot navigation, and drone flight. Although traditional path planning algorithms, such as the Dijkstra algorithm and the A* algorithm, can achieve good results in certain scenarios, they often have problems such as large amount of calculation, slow convergence, and easy to fall into local optimal solutions when facing complex and changing environments and large-scale data.

[0003] Ant Colony Optimization (ACO), as an optimization algorithm that simulates the foraging behavior of ants in nature, has been widely used in the field of path planning due to its good global optimization capabilities, inherent parallelism, and ease of computer implementation. In the ant colony algorithm, ants release a chemical called "pheromone" on the path they pass through while searching for food. Other ants can sense these pheromones and tend to move along the path with higher pheromone concentrations. This positive feedback mechanism enables the ants to eventually find the shortest path from the nest to the food source.

[0004] However, the traditional ant colony algorithm also has some obvious shortcomings in path planning. First, due to the fixed nature of the pheromone volatility factor, the algorithm is prone to search stagnation and premature convergence during the search process, that is, after the algorithm finds a relatively good path, it is difficult to jump out of the path for global search, thus limiting the algorithm's search ability and the diversity of solutions. Secondly, when choosing the next moving direction, the traditional ant colony algorithm mainly relies on the pheromone concentration and heuristic function at the current location, and the heuristic function often only considers the distance between the next two nodes to be moved, resulting in the ants lacking a clear goal in the process of movement, increasing the blindness and time-consuming nature of the search.

[0005] In order to solve these problems, researchers have proposed a variety of improved ant colony algorithms. For example, by introducing direction angles and bidirectional search strategies to improve the algorithm's optimization ability and convergence speed; combining genetic algorithms and other optimization algorithms to enhance the algorithm's global search ability; and adjusting algorithm parameters according to the characteristics of specific problems. These improved methods have improved the performance of the ant colony algorithm to a certain extent, but there are still some problems, such as complex algorithm parameter settings and large amount of calculations.

[0006] Therefore, this paper proposes a path planning method based on an improved ant colony algorithm, aiming to overcome the shortcomings of the traditional ant colony algorithm in path planning and improve the convergence speed and global search capability of the algorithm. Summary of the invention

[0007] The present invention proposes an improved ant colony algorithm to solve the path planning problem. By non-uniformly initializing the pheromone concentration, adding the distance from the candidate node to the target node into the heuristic function, and combining the enhanced direction guidance mechanism and diversified selection transfer strategies, the algorithm's global search and local optimality capabilities are enhanced.

[0008] In order to solve the existing challenges, the present invention is implemented through the following technical solutions:

[0009] A path planning method based on an improved ant colony algorithm comprises the following steps:

[0010] Step 1: Define parameters. Determine the key parameters such as the number of ants, pheromone volatility coefficient p, pheromone importance factor α, distance heuristic factor β, selection transfer strategy factor q, maximum number of iterations I, starting and end point position coordinates, and total pheromone amount Q.

[0011] Step 2: Initialization. Generate a colony of multiple virtual ants and place them all at the starting point. Initialize the pheromone concentration of each position non-uniformly. Specifically, the pheromone concentration of each position is initialized according to the sum of the Euclidean distances from each position to the starting point and the target point. The smaller the sum of the Euclidean distances, the higher the pheromone concentration, and vice versa. Initialize the taboo table to record the nodes that each ant has visited to prevent repeated visits.

[0012] Step 3, path selection. The coordinate axis is established with the starting node as the origin. The direction of the arrow is the direction from the starting node to the target node. The next optional node of all nodes should follow the direction of the arrow, that is, each node has at most three next candidate nodes, and the obstacle node is inaccessible. The optional nodes in these three directions are added to the neighbor set. Each ant selects the next node to be visited at a certain position based on the pheromone concentration between the current node and the candidate node and the inverse of the sum of the distances from the current node to the next node and the next node to the target node. The selection rule is based on pheromone concentration and distance heuristic information. In the early stage of the algorithm, a strategy combining roulette and deterministic transfer to the direction with the highest probability is used. When the random number q0 between 0 and 1 is greater than the selection transfer strategy factor q, the roulette strategy is used. Otherwise, the strategy of deterministic transfer to the direction with the highest probability is used. In the middle and late stages of the algorithm, a strategy combining the tournament method and the roulette method is used. When the random number q0 is greater than the selection transfer strategy factor q, the roulette strategy is used. Otherwise, the tournament strategy is selected.

[0013] Step 4: Path update and pheromone deposition. Each time an ant traverses a complete path, the path needs to be updated. When updating the path pheromone, the effect of the current path needs to be considered and evaluated. The pheromone on the path that the ant successfully reaches the target node increases, and the amount of increase is inversely correlated with the path length, that is, the shorter the path, the more pheromone is added. The pheromones on all paths will gradually evaporate over time, simulating the natural evaporation of pheromones over time.

[0014] Step 5: Determine whether the algorithm has reached the maximum number of iterations or meets the algorithm termination condition. If so, the algorithm outputs the optimal path found and the algorithm ends. Otherwise, execute the above steps again.

[0015] Compared with the prior art, the present invention has the following characteristics:

[0016] 1. The traditional ant colony algorithm assigns the same pheromone concentration to all paths during initialization, which results in the lack of clear goals in the movement of ants in the early stage of the algorithm, increasing the blindness and time-consuming nature of the search. The present invention initializes the pheromone concentration of each position non-uniformly according to the distance from each position to the starting node and the target node during initialization, thereby accelerating the convergence speed and optimization ability of the algorithm.

[0017] 2. The movable positions of the ants in the present invention are no longer the eight directions in the traditional algorithm, but the coordinate axis is established with the starting node as the origin, the arrow direction is the direction from the starting node to the target node, and the next optional nodes of all nodes should follow the direction of the arrow, that is, the next optional nodes of each node are at most three. This can reduce the calculation amount of the algorithm, enhance the direction guidance of the algorithm, reduce redundant search, and speed up the convergence speed of the algorithm.

[0018] 3. The present invention adopts a combination of multiple transfer strategies. In the early stage of the algorithm, a strategy combining random transfer using the roulette method and deterministic transfer to the direction with the highest probability is adopted, which can accelerate the convergence speed of the algorithm and enhance the algorithm's optimization ability. In the middle and late stages of the algorithm, a strategy combining the tournament method and the roulette method is adopted to expand the search range and help the algorithm jump out of the local optimum.

[0019] 4. The distance from the candidate node to the target point is added to the direction heuristic information, which improves the purposefulness and efficiency of path search. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Flowchart of path planning method based on improved ant colony algorithm

[0021] Figure 2 Schematic diagram of the direction guidance mechanism in the algorithm DETAILED DESCRIPTION

[0022] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in combination with specific examples and with reference to the accompanying drawings.

[0023] See also Figure 1 , a path planning method based on an improved ant colony algorithm, specifically comprising the following steps:

[0024] Step 1: Define some basic parameters, such as the number of ants M, pheromone volatility coefficient p, pheromone importance factor α, distance heuristic factor β, transfer strategy selection factor q, maximum number of iterations I, starting point and end point position coordinates, etc.

[0025] Step 2: Initialize. Generate a grid space that meets the conditions, such as Figure 2 As shown in the figure, S is the starting point, T is the target node, the black squares represent obstacles, and the white squares are the positions where the ants can move. Generate a group of C virtual ants and place them all at the starting point. Initialize the taboo table to record the nodes that each ant has visited to prevent repeated visits. Non-uniformly initialize the pheromone concentration of each position. Specifically, the pheromone concentration of each position is initialized according to the sum of the Euclidean distances from each position to the starting point and the target point. The smaller the sum of the Euclidean distances, the higher the pheromone concentration. Conversely, the pheromone concentration is lower, as shown below.

[0026]

[0027] where d min represents the Euclidean distance from the starting point to the end point, d Si and d iT Respectively represent the Euclidean distance from the current node to the starting point and the Euclidean distance from the current node to the target node, τ ij (initial) represents the initial pheromone concentration from position i to position j.

[0028] Step 3: The ants start searching and selecting paths. The coordinate axis is established with the starting node as the origin. The direction of the arrow is the direction from the starting node to the target node. The next optional nodes of all nodes should follow the direction of the arrow, such as Figure 2 As shown, S is the starting point, T is the target node, and the ant at position D can only move to D3, D4, and D5.

[0029] Add all optional nodes to the neighbor set. Each ant selects the next node to visit at a certain position based on the pheromone concentration between the current node and all candidate nodes and the inverse of the sum of the distances from the current node to the candidate node and the candidate node to the target node. The selection rule is based on pheromone concentration and distance heuristic information. The specific selection probability can be calculated by the formula, which is as follows

[0030]

[0031]

[0032] Among them, α and β are pheromone important factors and distance heuristic factors, represents the probability of ant m transferring from the i-th node to the j-th node, τ(i, j) represents the pheromone concentration on the path, and d ij and d jT They represent the Euclidean distance from the current node to the candidate node and the Euclidean distance from the candidate node to the target node, respectively. η(i, j) is the heuristic function, which represents the reciprocal of the sum of the distances from the current node to the candidate node and from the candidate node to the target node. J m Indicates the optional nodes for ant m's next move.

[0033] When the transfer probability of each node is calculated, the transfer strategy factor q is selected. In the early stage of the algorithm, a combination of the roulette method and deterministic transfer to the direction with the highest probability is used. When the random number q0 between 0 and 1 is greater than the transfer strategy factor q, the roulette strategy is used, that is, random transfer is performed based on the selection probability of each node. Otherwise, a deterministic transfer strategy to the direction with the highest probability is used. The specific formula is as follows:

[0034] j=argmax{[τ ij (t)] α [η ij (t)] β}, q <q0

[0035] Where j represents the position of the next node to be reached; argmax() is a function used to determine the position of the maximum value, q0 is a random number uniformly distributed in the range [0, 1], the parameter q is a constant between 0 and 1, and τ ij (t) represents the pheromone concentration at position j, [η ij (t)] is the inverse of the distance between the current node i and the next node j, and σ and β are the pheromone importance factor and distance heuristic factor.

[0036] In the middle and late stages of the algorithm, a strategy combining the binary tournament strategy and the roulette method is used. When the random number q0 between 0 and 1 is greater than the transfer strategy factor q, the roulette strategy is used. Otherwise, the tournament strategy is selected, that is, two movable positions are randomly selected, and the transfer is made to the position of the two positions that is closer to the target point.

[0037] Step 4: Each time an ant traverses a complete path, it needs to update the path. When updating the path pheromone, it is necessary to consider the effect of the current path and evaluate it. The pheromone on the path that the ant successfully reaches the target node increases, and the amount of increase is inversely correlated with the path length (that is, the shorter the path, the more pheromone is added). The specific calculation formula is as follows:

[0038]

[0039]

[0040] Where p is the pheromone volatility factor, is the original pheromone concentration at the current location, is the change in pheromone concentration of ant m, Q is the total amount of pheromone carried by the ant, L m represents the path length of ant m.

[0041] The pheromones on all paths will gradually evaporate over time, simulating the natural evaporation of pheromones over time.

[0042] Step 5: Determine whether the algorithm has reached the maximum number of iterations or meets the algorithm termination condition. If so, the algorithm outputs the optimal path found and the algorithm ends. Otherwise, the above steps are executed again. By comparing the optimal path obtained by the present invention with the optimal path obtained by the traditional ant colony algorithm, it can be concluded that the improved ant colony algorithm in the present invention can find a shorter and relatively smoother optimal path more quickly.

[0043] It should be noted that although the embodiments of the present invention described above are illustrative, they are not intended to limit the present invention, and therefore the present invention is not limited to the above specific embodiments. Without departing from the principles of the present invention, any other embodiments obtained by those skilled in the art under the guidance of the present invention are deemed to be within the protection of the present invention.

Claims

1. A path planning method based on an improved ant colony algorithm comprises the following steps: Step 1: Define parameters. Determine the key parameters such as the number of ants, pheromone volatility coefficient p, pheromone importance factor α, distance heuristic factor β, selection transfer strategy factor q, maximum number of iterations I, starting and end point position coordinates, and total pheromone amount Q. Step 2: Initialization. Generate a colony of multiple virtual ants and place them all at the starting point. Non-uniformly initialize the pheromone concentration at each location. Specifically, the pheromone concentration at each location is initialized based on the sum of the Euclidean distances from each location to the starting point and the target point. The smaller the sum of the Euclidean distances, the higher the pheromone concentration. Conversely, the lower the pheromone concentration. Initialize the taboo table to record the nodes that each ant has visited to prevent repeated visits. Step 3: Path selection. The coordinate axis is established with the starting node as the origin. The direction of the arrow is the direction from the starting node to the target node. The next optional nodes of all nodes should follow the direction of the arrow, that is, each node has at most three next candidate nodes. Obstacle nodes are inaccessible, and the optional nodes in these three directions are added to the neighbor set. Each ant selects the next node to be visited at a certain position based on the pheromone concentration between the current node and the candidate node and the inverse of the sum of the distances from the current node to the candidate node and the candidate node to the target node. The selection rule is based on pheromone concentration and distance heuristic information. According to the selection transfer strategy factor q, in the early stage of the algorithm, a strategy combining the roulette method and deterministic transfer to the direction with the highest probability is used. When the randomly generated random number q0 between 0 and 1 is greater than the selection transfer strategy factor q, the roulette strategy is used. Otherwise, the strategy of deterministic transfer to the direction with the highest probability is used. In the middle and late stages of the algorithm, a strategy combining the tournament method and the roulette method is used. When the random number q0 is greater than the selection transfer strategy factor q, the roulette strategy is used. Otherwise, the tournament strategy is selected. Step 4: Path update and pheromone deposition. Each time an ant traverses a complete path, the path needs to be updated. When updating the path pheromone, the effect of the current path needs to be considered and evaluated. The pheromone on the path that the ant successfully reaches the target node increases, and the amount of increase is inversely correlated with the path length, that is, the shorter the path, the more pheromone is added. The pheromones on all paths will gradually evaporate over time, simulating the natural evaporation of pheromones over time. Step 5: Determine whether the algorithm has reached the maximum number of iterations or meets the algorithm termination condition. If so, the algorithm outputs the optimal path found and the algorithm ends. Otherwise, execute the above steps again.

2. According to the path planning method based on the improved ant colony algorithm of claim 1, the pheromone concentration of each position is initialized non-uniformly in step 2, as shown below: where d min represents the Euclidean distance from the starting point to the end point, d Si and d iT Respectively represent the Euclidean distance from the current node to the starting point and the Euclidean distance from the current node to the target node, τ ij (initial) represents the initial pheromone concentration from position i to position j.

3. According to the path planning method based on the improved ant colony algorithm of claim 1, the specific calculation method of the selection probability in step 3 is as shown in the following formula: Among them, α and β are pheromone important factors and distance heuristic factors, represents the probability of ant m transferring from the i-th node to the j-th node, τ(i, j) represents the pheromone concentration on the path, and d ij and d jT They represent the Euclidean distance from the current node to the candidate node and the Euclidean distance from the candidate node to the target node, respectively. η(i, j) is the heuristic function, which represents the reciprocal of the sum of the distances from the current node to the candidate node and from the candidate node to the target node. J m Indicates the optional nodes for ant m's next move.

4. According to the path planning method based on the improved ant colony algorithm of claim 1, the strategy of deterministically transferring to the direction with the highest probability in step 3 is specifically as follows: j=argmax{[τ ij (t)] α [η ij (t)] β },q<q0 Where j represents the position of the next node to be reached, argmax() is a function used to determine the position of the maximum value, q0 is a random number in the range of [0, 1], and the parameter q is a constant between 0 and 1. The tournament strategy in step 3 specifically refers to randomly selecting two nodes from the optional nodes, and the ants move to the node that is closer to the target point.

5. According to the path planning method based on the improved ant colony algorithm of claim 1, the updating and volatilization process of pheromone in step 4 is specifically shown in the following formula: Where p is the pheromone volatility factor, is the original pheromone concentration at the current location, is the change in pheromone concentration of ant m, Q is the total amount of pheromone carried by the ant, L m represents the path length of ant m.

Images