阅读说明：本技术 一种基于均匀设计的路径规划蚁群算法参数选取方法 (Path planning ant colony algorithm parameter selection method based on uniform design ) 是由 史晓军 姚鑫 梅雪松 胡佳祥 王迎新 于 2021-06-29 设计创作，主要内容包括：本发明公开了一种基于均匀设计的路径规划蚁群算法参数选取方法,属于智能机器人路径规划领域。包括：选择路径规划中对蚁群算法的性能有影响的各因素的水平范围、水平及水平数；根据各因素的因素数及各因素的水平数,选择均匀设计表；将各因素的各水平填入均匀设计表中,得到实际试验参数表；基于得到的实际试验参数表中的每一组试验参数代入蚁群算法中进行路径规划,得到每一组试验参数的路径规划结果；对每一组试验参数的路径规划结果中对应的路径长度和迭代次数进行分析,得到所有试验次数中最优的蚁群算法参数组合。因此,本发明能够解决现有的蚁群算法的初始参数值只能通过经验来设置,而难以选取合适的参数组合问题。(The invention discloses a path planning ant colony algorithm parameter selection method based on uniform design, and belongs to the field of intelligent robot path planning. The method comprises the following steps: selecting the level range, level and level number of each factor which has influence on the performance of the ant colony algorithm in path planning; selecting a uniform design table according to the number of factors and the level number of the factors; filling each level of each factor into a uniform design table to obtain an actual test parameter table; substituting each group of test parameters in the obtained actual test parameter table into the ant colony algorithm for path planning to obtain a path planning result of each group of test parameters; and analyzing the path length and the iteration times corresponding to the path planning result of each group of test parameters to obtain the optimal ant colony algorithm parameter combination in all the test times. Therefore, the method can solve the problem that the initial parameter value of the existing ant colony algorithm can only be set through experience, and is difficult to select a proper parameter combination.)

1. A path planning ant colony algorithm parameter selection method based on uniform design is characterized by comprising the following steps:

1) selecting each factor which has influence on the performance of the ant colony algorithm in path planning, and determining the horizontal range, the horizontal level and the horizontal level number of each factor;

2) selecting a uniform design table according to the number of factors and the level number of the factors; filling each level of each factor into a uniform design table to obtain an actual test parameter table;

3) substituting each group of test parameters in the obtained actual test parameter table into the ant colony algorithm for path planning to obtain a path planning result of each group of test parameters; and analyzing the path length and the iteration times corresponding to the path planning result of each group of test parameters to obtain the optimal ant colony algorithm parameter combination in all the test times.

2. The method for selecting the ant colony algorithm parameters for path planning based on uniform design according to claim 1, wherein in the step 1), the horizontal range, the horizontal number and the horizontal level of each factor are determined, and the specific operations comprise: determining the horizontal range of each factor according to the common parameter range of the ant colony algorithm, setting each level according to the test times and the precision or scale of the parameters in the horizontal range, and obtaining the level number according to the number of the levels.

3. The method for selecting parameters of the ant colony algorithm for path planning based on uniform design according to claim 1, wherein in the step 2), the uniform design table is as follows: u shape_n(m^r)；

Wherein n represents the number of uniform design table rows, namely the required test times; m represents the level number of each factor; r represents the number of columns of the uniform design table, i.e., the number of factors that can be accommodated.

4. The method for selecting the ant colony algorithm parameters for path planning based on uniform design according to claim 1, wherein in the step 3), each set of test parameters in the actual test parameter table is substituted into the ant colony algorithm for path planning, and the specific steps are as follows:

firstly, modeling a working environment by adopting a grid method, and establishing a grid map;

secondly, initializing various parameters; wherein each parameter comprises a starting point and an end point of path planning, iteration times and all factors in the step 1);

creating and initializing a tabu table, and adding a starting point into the tabu table;

fourthly, the ants are put to the starting point to release the ants;

selecting the next node according to the transition probability formula and moving;

recording the path and adding the current node into a tabu table;

when the ants are judged to reach the end point or have no way to walk, continuing the next step:

eighty percent, adding 1 to the serial number of the ant;

when the ants are judged to be completely released, continuing the next step;

ninthly, updating the pheromone, recording the optimal path, and initializing the ants;

the ant initialization operation comprises the following steps: resetting the ant number, and adding 1 to the iteration number;

in the third step, when all the iteration times are judged to be finished, the next step is continued;

and outputting the global optimal path and the iteration curve in the grid map.

5. The method for selecting ant colony algorithm parameters for path planning based on uniform design according to claim 4, wherein in the fifth step, the transition probability formula is as follows:

in the formula, alpha is an information elicitation factor; β is the desired heuristic;the transfer probability of the ant k from the node i to the node j at the moment t; c is a set of the ant k going to the next feasible node from the current node, and when the algorithm runs, the information is stored in a tabu table; tau is_ij(t) the pheromone concentration of the paths from the node i to the node j at the time t, wherein the pheromone concentration of each path at the initial time is the same; eta_ij(t) a heuristic function, η, representing the nodes i to j at time t_ij(t)＝1/d_ijWherein d is_ijIs the euclidean distance from node i to node j.

6. The method as claimed in claim 4, wherein in step (c), when it is determined that ants do not reach the end point or are not allowed to travel without any way, step (c) is re-executed until the end.

7. The method for selecting parameters of an ant colony algorithm for path planning based on uniform design according to claim 4, wherein in step ((b)), when it is judged that ants are not completely released, the step (c) is executed again until the end.

8. The method for selecting ant colony algorithm parameters for path planning based on uniform design according to claim 4, wherein in the step ninthly, information is updated and an optimal path is recorded through the following formula;

τ_ij(t+1)＝(1-ρ)τ_ij(t)+Δτ_ij(t)；

in the formula, rho is a pheromone volatilization factor; m is the number of ants; tau is_ij(t +1) is the pheromone concentration after updating from node i to node j; tau is_ij(t) is the current pheromone concentration; delta tau_ij(t) is the sum of the pheromone concentrations accumulated by all ants passing through the path in the searching process;for pheromones left by the kth ant when the kth ant passes through a path from a node i to a node j, the pheromone updating model adopts the following formula:

wherein Q is a pheromone intensity factor; l is_kThe total path length of ant k after successfully reaching the end point.

9. The method for selecting parameters of ant colony algorithm for path planning based on uniform design according to claim 4 is characterized in that in step (c), when all iterations are not completed, step (c) is re-executed until the end.

10. The method for selecting parameters of an ant colony algorithm for path planning based on uniform design according to claim 1, wherein in the step 1), each factor comprises at least two of an information elicitation factor α, an expected elicitation factor β, an pheromone volatilization factor ρ, the number of ants m, and an pheromone intensity factor Q.

Technical Field

The invention belongs to the field of intelligent robot path planning, and relates to a path planning ant colony algorithm parameter selection method based on uniform design.

Background

The uniform design is a test design method which combines the number theory and the multivariate statistics by professor of kelita and royal of scientists and is jointly proposed in 1978 only by considering that test points are uniformly distributed in a test range. The test method focuses on obtaining the most information with the least number of tests in the test range, so that the test times are obviously reduced compared with the orthogonal design, and the advantage of uniform design can be better embodied when some factors have main effects or mutual effects. For example, when there are a total of m factors in a given trial, each factor having n levels, if the trial is performed using enumeration, the total number of trials is n^mNext, the process is carried out. If orthogonal design is used, the total number of trials is n²Next, the process is carried out. While the total number of trials using the uniform design was n. Therefore, the test times can be greatly reduced by uniform design. The uniform design table is a normalized array table with uniformity constructed according to the application principle of number theory in multi-dimensional numerical integration, is used for uniform design, and is the basis of uniform design.

The robot path planning technology is an important part in the research field of mobile robots and also an important link in robot navigation, and is a comprehensive index based on certain specific performance parameters (such as path planning length, time spent in the planning process, path tortuosity and the like, and most commonly, the path planning length), and a feasible optimal or suboptimal path without collision is searched from a starting point to a target point by using a path planning algorithm on the premise that obstacles are distributed in the environment. Researchers in various countries around the world in the past decades have researched many path planning algorithms, such as Dijkstra algorithm, a-algorithm and other traditional algorithms, and recently developed rapid intelligent bionic algorithms, such as particle swarm algorithm, ant colony algorithm, genetic algorithm and the like.

The ant colony algorithm is a meta-heuristic random search algorithm inspired by the foraging behavior of ants. The algorithm has the advantages of good robustness, strong global search capability, convenient environmental constraint expression and the like, and is widely applied to the combined optimization problem and the path planning of the mobile robot. The performance of the ant colony algorithm is determined by a plurality of parameters, and the main parameters comprise: information elicitation factor α, which represents the importance of pheromones on the path. The larger the value is, the weaker the randomness of the ant search is, the faster the convergence speed is, and the higher the probability of the path taken by the ant before selecting; the smaller the value is, the stronger the search randomness of the ant is, the slower the convergence speed is, and the ant is likely to fall into local optimum. Secondly, an elicitation factor beta is expected, the parameter represents the importance of visibility, and the size of the elicitation factor beta reflects the action strength of the priori and deterministic factors in the ant optimizing process. The larger the value is, the higher the possibility that ants select the local shortest path on the node is, and although the convergence speed of the algorithm is accelerated, the randomness of path search is weakened, and the ant is easy to fall into local optimum. The smaller the value is, the stronger the search randomness of the ant is, and the optimal solution is difficult to find. And the pheromone volatilization factor rho represents the volatilization degree of the pheromone on the path. The corresponding (1-rho) is an pheromone residual factor, represents the residual degree of pheromones on the path, and represents the information exchange strength between ants. If the rho value is larger, the influence time of the pheromone is shorter, so that the pheromone on the path which is not searched for in any way approaches to 0, local optimization is easy to be trapped, and the global searching capability and the randomness of the algorithm are reduced. The smaller the rho value is, the randomness and the global search capability of the algorithm can be improved, but the convergence rate of the algorithm is reduced. And the larger the number m of the ants is due to the parallelism of the ant colony algorithm, the stronger the global searching capability of the algorithm and the stability of the algorithm are, but the distribution difference of the pheromone is weakened, the influence of the pheromone is reduced, and the convergence speed is reduced. On the contrary, although the convergence rate is accelerated, the algorithm stability and the global search capability are weaker. And fifthly, the pheromone intensity factor Q represents the total amount of pheromone released by the ant on the path after the ant completes the path searching process. The larger the value of the algorithm is, the higher the convergence of the algorithm is, but the global searching capability of the algorithm is weakened, so that the algorithm is more prone to fall into local optimum. Therefore, each parameter of the ant colony algorithm influences the global searching capability and the convergence speed of the algorithm. The invention provides a path planning ant colony algorithm parameter selection method based on uniform design, which can obtain excellent parameter combinations through a small amount of tests and greatly improve the conditions.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention aims to provide a method for selecting parameters of a path planning ant colony algorithm based on uniform design, and solves the problems that initial parameter values of the ant colony algorithm in the prior art can only be set through experience and a proper parameter combination is difficult to select efficiently.

In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:

the invention discloses a path planning ant colony algorithm parameter selection method based on uniform design, which comprises the following steps:

1) selecting each factor which has influence on the performance of the ant colony algorithm in path planning, and determining the horizontal range, the horizontal level and the horizontal level number of each factor;

2) selecting a uniform design table according to the number of factors and the level number of the factors; filling each level of each factor into a uniform design table to obtain an actual test parameter table;

3) substituting each group of test parameters in the obtained actual test parameter table into the ant colony algorithm for path planning to obtain a path planning result of each group of test parameters; and analyzing the path length and the iteration times corresponding to the path planning result of each group of test parameters to obtain the optimal ant colony algorithm parameter combination in all the test times.

Preferably, in step 1), the level range, level and level number of each factor are determined, and the specific operations include: determining the horizontal range of each factor according to the common parameter range of the ant colony algorithm, setting each level according to the test times and the precision or scale of the parameters in the horizontal range, and obtaining the level number according to the number of the levels.

Preferably, in step 2), the uniform design table is: u shape_n(m^r)；

Wherein n represents the number of uniform design table rows, namely the required test times; m represents the level number of each factor; r represents the number of columns of the uniform design table, i.e., the number of factors that can be accommodated.

Preferably, in step 3), each set of test parameters in the actual test parameter table is substituted into the ant colony algorithm for path planning, and the specific steps are as follows:

firstly, modeling a working environment by adopting a grid method, and establishing a grid map;

secondly, initializing various parameters; wherein each parameter comprises a starting point and an end point of path planning, iteration times and all factors in the step 1);

creating and initializing a tabu table, and adding a starting point into the tabu table;

fourthly, the ants are put to the starting point to release the ants;

selecting the next node according to the transition probability formula and moving;

recording the path and adding the current node into a tabu table;

when the ants are judged to reach the end point or have no way to walk, continuing the next step:

eighty percent, adding 1 to the serial number of the ant;

when the ants are judged to be completely released, continuing the next step;

ninthly, updating the pheromone, recording the optimal path, and initializing the ants;

the ant initialization operation comprises the following steps: resetting the ant number, and adding 1 to the iteration number;

in the third step, when all the iteration times are judged to be finished, the next step is continued;

outputting the global optimal path and the iteration curve in the grid map.

Further preferably, in the fifth step, the transition probability formula is:

in the formula, alpha is an information elicitation factor; β is the desired heuristic;the transfer probability of the ant k from the node i to the node j at the moment t; c is a set of the ant k going to the next feasible node from the current node, and when the algorithm runs, the information is stored in a tabu table; tau is_ij(t) the pheromone concentration of the paths from the node i to the node j at the time t, wherein the pheromone concentration of each path at the initial time is the same; eta_ij(t) a heuristic function, η, representing the nodes i to j at time t_ij(t)＝1/d_ijWherein d is_ijIs the euclidean distance from node i to node j.

Further preferably, in step (c), when it is determined that the ants do not reach the end point or are not available, re-executing step (c) until the end.

Further preferably, in the step ((b)), when the ant is not completely released, the step ((c)) is executed again until the end.

Further preferably, in the step ninthly, the information is updated and the optimal path is recorded through the following formula;

τ_ij(t+1)＝(1-ρ)τ_ij(t)+Δτ_ij(t)；

in the formula, rho is a pheromone volatilization factor; m is the number of ants; tau is_ij(t +1) is the pheromone concentration after updating from node i to node j; tau is_ij(t) is the current pheromone concentration; delta tau_ij(t) is the sum of the pheromone concentrations accumulated by all ants passing through the path in the searching process;the signal left by the kth ant when passing through the path from node i to node jThe pheromone updating model adopts the following formula:

wherein Q is a pheromone intensity factor; l is_kThe total path length of ant k after successfully reaching the end point.

Further preferably, in step r, when it is judged that all iterations have not been completed, step c is re-executed until the end.

Preferably, in step 1), each factor includes at least two of an information elicitation factor α, an expected elicitation factor β, a pheromone volatilization factor ρ, the number of ants m, and a pheromone intensity factor Q.

Compared with the prior art, the invention has the following beneficial effects:

the invention discloses a path planning ant colony algorithm parameter selection method based on uniform design, which can obtain excellent ant colony algorithm parameter combination by utilizing a small amount of tests performed by the uniform design, so that the ant colony algorithm can obtain shorter path length, relatively fewer iteration times and higher convergence rate. The method of the invention utilizes the efficient parameter selection method to obtain excellent parameter combinations at the beginning of the operation of the algorithm, and is very beneficial to the subsequent operation or online adjustment of the algorithm. In practical application, a lot of time can be saved in the parameter selection of the ant colony algorithm in the path planning. Therefore, the method can effectively solve the problem that the initial parameter value of the ant colony algorithm can only be set through experience and is difficult to select a proper parameter combination.

Drawings

FIG. 1 is a flow chart of a method of the present invention;

FIG. 2 is a flow chart of an ant colony algorithm used in the present invention;

FIG. 3 is a schematic diagram of a grid environment map in the present embodiment; wherein, (a) is a simple grid map, and (b) is a complex grid map;

FIG. 4 is a diagram illustrating a path planning result obtained by the present invention; wherein, (a) is a path obtained under a simple grid map, and (b) is a path obtained under a complex grid map;

FIG. 5 is a graph illustrating the trend of the convergence curve obtained by the present invention; wherein, (a) is the change trend of the convergence curve obtained under a simple grid map, and (b) is the change trend of the convergence curve obtained under a complex grid map;

FIG. 6 is a diagram illustrating a result distribution graph selected by using enumeration parameters according to the present embodiment; the map is obtained by using an enumeration method under a simple map, and the map is obtained by using an enumeration method under a complex map.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

The invention is described in detail below with reference to the accompanying drawings:

referring to fig. 1, the invention discloses a method for selecting ant colony algorithm parameters based on path planning of uniform design, which comprises the following steps:

step1, selecting factors which have influence on the performance of the ant colony algorithm in path planning (the factors comprise at least two of an information elicitation factor alpha, an expected elicitation factor beta, an pheromone volatilization factor rho, the number m of ants and an pheromone intensity factor Q);

step2, determining the horizontal range of each factor selected by Step1 according to the common parameter range of the ant colony algorithm to obtain the value range of each factor, and setting each level according to the test times and the precision or scale of the parameters in the horizontal range, wherein the number of the levels is the horizontal number;

specifically, in one embodiment of the present invention, the number of levels generally depends on the number of trials and the precision or scale of the parameter, such as α ∈ (1,10), and α must be an integer, so the number of levels of α is taken to be at most 10, and the corresponding level is (1,2,3,4,5,6,7,8,9, 10); if taken as 20, the corresponding levels are (1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10, 10). Here, the parameters of the ant colony algorithm mainly depend on the number of trials.

Here, there is no precision limit for the ant colony algorithm, and it mainly depends on the number of trials.

Step3 selecting uniform design Table U based on the number of factors determined at Step1 and the number of levels of factors determined at Step2_n(m^r) Wherein n represents the number of uniform design table rows, i.e. the number of required tests; m represents the level number of each factor; r represents the number of columns of the uniform design table, i.e., the number of factors that can be accommodated;

step4, filling each level of each factor determined in Step2 into the corresponding position in the uniform design table obtained in Step3 to obtain an actual test parameter table;

step5, substituting each group of test parameters in all test times in the actual test parameter table obtained in Step4 into the ant colony algorithm for path planning to obtain a path planning result of each group of test parameters;

step6, analyzing the path length and the iteration times corresponding to the path planning result of each group of test parameters;

and Step7, obtaining the optimal ant colony algorithm parameter combination in all test times.

According to the method, the method for selecting the ant colony algorithm parameters based on the path planning of the uniform design is realized according to steps 1-7.

Referring to fig. 2, for each set of test parameters in all test times in the actual test parameter table described in Step5 above to perform path planning by substituting the test parameters into the ant colony algorithm, the specific steps and flows are as follows:

step5.1, modeling the working environment of the robot by adopting a grid method, and establishing a grid map;

step5.2, initializing each parameter of a starting point and an end point of path planning, an information heuristic factor alpha, an expected heuristic factor beta, an pheromone volatilization factor rho, the number m of ants, an pheromone intensity factor Q and the iteration number based on the established grid map, and establishing an adjacent matrix, a heuristic information matrix and an initial pheromone matrix of the grid map;

step5.3, establishing and initializing a tabu table, and adding a starting point into the tabu table;

step5.4, placing the ants to the starting point, and releasing the ants;

step5.5, selecting a next node according to a transition probability formula (1) and moving;

in the formula (1), alpha is an information elicitation factor; β is the desired heuristic;the transfer probability of the ant k from the node i to the node j at the moment t; c is a set of the ant k going to the next feasible node from the current node, and when the algorithm runs, the information is stored in a tabu table; tau is_ij(t) represents the pheromone concentration of the path from node i to node j at time t, wherein the initial time isPheromone concentrations of all paths are the same; eta_ij(t) denotes the heuristic function of node i to node j at time t, also called visibility, usually taken as η_ij(t)＝1/d_ijWherein d is_ijThe Euclidean distance from the node i to the node j;

step5.6, recording the path, and adding the current node into a tabu table;

and Step5.7, judging whether the ants reach the terminal or can move without a road currently. If yes, continuing the next step, otherwise jumping to Step5.5;

step5.8, adding 1 to ant number, and judging whether ants are all released. If yes, continuing the next step, otherwise jumping to Step5.3;

step5.9, updating pheromone according to formulas (2) and (3), recording the optimal path, setting the ant number to be 1, and adding 1 to the iteration number;

τ_ij(t+1)＝(1-ρ)τ_ij(t)+Δτ_ij(t) (2)

in the formula, rho is a pheromone volatilization factor; m is the number of ants; tau is_ij(t +1) is the pheromone concentration after updating from node i to node j; tau is_ij(t) is the current pheromone concentration; delta tau_ij(t) is the sum of the pheromone concentrations accumulated by all ants that pass through the path during the search.For pheromones left by the kth ant when the kth ant passes through a path from a node i to a node j, a periant model is adopted for the pheromone updating model as shown in a formula (4);

wherein Q is a pheromone intensity factor; l is_kThe total path length of ant k after successfully reaching the end point. Obviously, the shorter the path the ant seeks, the higher the concentration of pheromones, and thusTo facilitate the search for the optimal path;

and Step5.10, judging whether all the iteration times are finished. If yes, continuing the next step, otherwise jumping to Step5.3;

and Step5.11, outputting the global optimal path in the grid map and outputting an iteration curve.

The invention is described in further detail below with reference to the following figures and specific examples:

examples

A path planning ant colony algorithm parameter selection method based on uniform design comprises the following steps:

step1 five factors in the choice path planning that have an impact on the performance of the ant colony algorithm: information elicitation factor alpha, expected elicitation factor beta, pheromone volatilization factor rho, ant number m and pheromone intensity factor Q;

step2, determining the horizontal ranges alpha (1-9), beta (1-9), rho (0.1-0.9), m (20-60) and Q (1-80) of all factors according to the common parameter range of the ant colony algorithm, and then taking a proper level (value) in the horizontal range;

step3 selecting a uniform design table U based on the number of factors and the number of levels₁₅(15⁵) Wherein the number of uniformly designed table rows is 15, namely the required test times; factor level number 15; the column number of the uniform design table is 5, namely the number of factors which can be accommodated;

step4, filling the levels of each factor into the corresponding positions in the uniform design table to obtain the actual test parameter table shown in table 1:

TABLE 1 Uniform design Table and test parameters

Step5, substituting each group of test parameters in the actual test parameter table into the ant colony algorithm for path planning;

step6, analyzing the path length and the iteration times corresponding to each group of results to obtain a test group with the best effect, wherein the path length corresponding to the test group is shown in FIG. 4, and a convergence curve trend change chart is shown in FIG. 5;

step7 the optimal ant colony algorithm parameter combinations in these experiments were obtained.

For the ant colony algorithm path planning described in Step5, the specific steps are as follows:

step5.1, modeling the working environment of the robot by adopting a grid method, and establishing a grid map, wherein a white grid is a free grid and represents an area where the robot can pass, and a black grid is an obstacle grid and represents an area where the robot cannot pass, as shown in figure 3. In the whole path planning process, all information in the map is fixed and unchanged, a robot with the volume smaller than a unit grid is regarded as a particle, a Cartesian coordinate system is established in a grid environment, each grid is sequentially numbered from left to right and from top to bottom, and then the center coordinate (x) of each grid is obtained_i,y_i) The mapping relation with the grid number is formula (5):

in the formula: i is a grid number; m is the width of the grid map; n is the length of the grid map; mod is a remainder operation and fix is a rounding operation;

step5.2, initializing each parameter of a starting point and an end point of path planning, an information elicitation factor alpha, an expected elicitation factor beta, an pheromone volatilization factor rho, the number m of ants, an pheromone intensity factor Q and the iteration number, and establishing an adjacent matrix, a heuristic information matrix and an initial pheromone matrix of the grid map;

step5.3, establishing and initializing a tabu table, and adding a starting point into the tabu table;

step5.4, placing the ants to the starting point, and releasing the ants;

step5.5, selecting a next node and moving the next node according to a transition probability formula (1) based on an adjacent matrix, a heuristic information matrix and an initial pheromone matrix of the grid map;

in the formula (1), alpha is an information elicitation factor; β is the desired heuristic;the transfer probability of the ant k from the node i to the node j at the moment t; c is a set of the ant k going to the next feasible node from the current node, and when the algorithm runs, the information is stored in a tabu table; tau is_ij(t) the pheromone concentration of the paths from the node i to the node j at the time t, wherein the pheromone concentration of each path at the initial time is the same; eta_ij(t) denotes the heuristic function of node i to node j at time t, also called visibility, usually taken as η_ij(t)＝1/d_ijWherein d is_ijThe Euclidean distance from the node i to the node j;

step5.6, recording the path, and adding the current node into a tabu table;

and Step5.7, judging whether the ants reach the terminal or can move without a road currently. If yes, continuing the next step, otherwise jumping to Step5.5;

step5.8, adding 1 to ant number, and judging whether ants are all released. If yes, continuing the next step, otherwise jumping to Step5.3;

step5.9, updating pheromones and recording an optimal path according to formulas (2) and (3) based on an adjacent matrix, a heuristic information matrix and an initial pheromone matrix of the grid map, setting the ant number to be 1, and adding 1 to the iteration number;

τ_ij(t+1)＝(1-ρ)τ_ij(t)+Δτ_ij(t) (2)

wherein, tau_ij(t +1) is the pheromone concentration after updating from node i to node j; tau is_ij(t) is the current pheromone concentration; delta tau_ij(t) is the sum of the pheromone concentrations accumulated by all ants that pass through the path during the search.For pheromones left by the kth ant when the kth ant passes through a path from a node i to a node j, a periant model is adopted for the pheromone updating model as shown in a formula (4);

wherein L is_kThe total path length of ant k after successfully reaching the end point. Obviously, the shorter the path sought by the ant, the higher the concentration of pheromones, facilitating the search for the optimal path.

And Step5.10, judging whether all the iteration times are finished. If yes, continuing the next step, otherwise jumping to Step5.3;

and Step5.11, outputting the global optimal path in the grid map and outputting an iteration curve.

In order to show the advantages of the present invention, an enumeration method experiment is performed on the ant colony algorithm parameter range in the embodiment, and a parameter selection result distribution diagram of the enumeration method is obtained as shown in fig. 6, which can obtain: the main parameters influencing the performance of the ant colony algorithm are an information heuristic factor alpha and an expected heuristic factor beta, and the results obtained by two maps with different complexity degrees are approximately the same and have consistent trends.

The optimal parameter combination obtained by using the parameter selection method based on uniform design in the invention is in the optimal result range in the enumeration method. With the path length as an evaluation index, for the two maps in the embodiment, the result of obtaining the optimal parameter combination by using uniform design can be ranked by about 10% in the effective result of the enumeration method. In addition, the iteration times of the algorithm are used as evaluation indexes, the iteration times of the ant colony algorithm combined by using the optimal parameters obtained by uniform design are very small, and the convergence speed is very high. Therefore, in the embodiment, it can be seen that the parameter selection using the uniform design method has a very obvious advantage compared with the huge workload of the enumeration method.

When the ant colony algorithm is actually used for path planning, the uniform design method not only can efficiently obtain excellent parameter combinations while considering a plurality of parameters, but also can avoid the blindness and the inefficiency of methods such as direct parameter value taking or common parameter combination under the common condition.

In summary, the present invention relates to a method for selecting ant colony algorithm parameters for path planning based on uniform design. Since the performance of the ant colony algorithm is determined by a plurality of parameters, each parameter affects the global searching capability and the convergence speed of the algorithm, and the relationship among the parameters is close. The method selects parameters of the ant colony algorithm by utilizing uniform design, and obtains high-quality parameter combinations through a small amount of experiments. In the robot path planning, the method can quickly obtain better parameter combinations, so that the ant colony algorithm has better performance and can be converged to the global optimal solution more quickly.

The above-mentioned contents are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modification made on the basis of the technical idea of the present invention falls within the protection scope of the claims of the present invention.