CN115167460A - Robot cluster task optimal path planning method based on fusion algorithm - Google Patents
Robot cluster task optimal path planning method based on fusion algorithm Download PDFInfo
- Publication number
- CN115167460A CN115167460A CN202210941660.4A CN202210941660A CN115167460A CN 115167460 A CN115167460 A CN 115167460A CN 202210941660 A CN202210941660 A CN 202210941660A CN 115167460 A CN115167460 A CN 115167460A
- Authority
- CN
- China
- Prior art keywords
- robot
- path
- sub
- task target
- target point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 54
- 230000004927 fusion Effects 0.000 title claims abstract description 16
- 239000003016 pheromone Substances 0.000 claims abstract description 56
- 230000002068 genetic effect Effects 0.000 claims abstract description 12
- 238000012546 transfer Methods 0.000 claims description 37
- 238000012216 screening Methods 0.000 claims description 8
- 238000012163 sequencing technique Methods 0.000 claims description 8
- 230000007704 transition Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 6
- 230000035772 mutation Effects 0.000 claims description 3
- 108091026890 Coding region Proteins 0.000 claims description 2
- 238000010353 genetic engineering Methods 0.000 claims 1
- 230000007547 defect Effects 0.000 description 6
- 238000005457 optimization Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 230000007123 defense Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0219—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory ensuring the processing of the whole working surface
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses a robot cluster task optimal path planning method based on a fusion algorithm. The method comprises the following implementation steps: generating an initial robot cluster path scheme by using a genetic algorithm; updating the task target point of each robot; updating the task target point of each robot by using a local search method; updating the final pheromone concentration by using an elite strategy; judging whether the updating times reach the optimal fitting updating times or not; judging whether the difference value between the total distance cost after each update and the total distance cost after the last update is less than 3% of the total distance cost after the iteration update after 5 times of continuous updates; and decoding to generate a sub-path of each robot as a robot cluster task optimal path planning scheme. Compared with the traditional ant colony algorithm, the method has the advantages of high convergence speed, high solving speed and difficulty in falling into local optimum.
Description
Technical Field
The invention belongs to the technical field of physics, and further relates to a robot cluster task optimal path planning method based on a fusion algorithm in the technical field of robot operation. The method can be used for planning the optimal path of the task executed by each robot in the robot cluster.
Background
A robot cluster is a group consisting of a plurality of robots intended to accomplish the same type of task. The tasks of the same type comprise a plurality of task target points, and different distance cost or time cost is needed for the robot to reach each task target point. For example, in a warehouse logistics environment, the logistics robot needs to transport goods to different warehouses; in geophysical surveying, a geo-robot needs to survey different sites. The system reasonably distributes tasks to each robot to expect that the total cost of the robot cluster is minimized in the process of executing the tasks. The system optimization is related to the quality and the working capacity of the whole cluster system and is an important ring in the initial work. At present, algorithms for performing optimal path planning on robot cluster tasks are mainly divided into two types: market-based methods and heuristic-based methods. The market-based method, also referred to as an auction-based method, may be configured such that bids are collectively received by a "central auctioneer" and task targets are assigned to the least expensive robot, or may be distributed among different robots to share the task targets. The heuristic algorithm is an algorithm based on intuitive or empirical construction, and gives a feasible solution to each instance of the combinatorial optimization problem to be solved at an acceptable cost.
The national defense science and technology innovation research institute of the military science institute of the people's liberation army of China discloses a method for planning paths of multiple robots based on a market algorithm in a patent document applied by the research institute of the national defense science and technology ' a method for rapidly establishing a collaborative map of multiple robots based on an improved market method ' (application number: CN202111252038, application publication number: CN 114137955A). The method is divided into two stages. The first stage is as follows: firstly, randomly selecting a robot and generating an initial main map; then, iteratively updating the initial main map; and forming all task combinations by using boundary points between the known area and the unknown area in the map, and adding the task combinations into the task set to be auctioned. And a second stage: the robot traverses all tasks in the task set, sequences the tasks obtained by auction to obtain a task list, and sets a first task in the task list as a current target point to advance; when the current target point is reached, the environment is scanned, and the process is circulated. This method has the disadvantage of being more complex since it generates a large number of possible combinations in the first stage and traverses them in the second stage. When a large-scale example is applied, the disadvantage of slow solving speed is brought by too high complexity.
The patent document "a multi-robot path planning method based on ant colony algorithm" (application number: CN 201910636641, application publication number: CN 110375759A) applied by Tianjin university discloses a method for performing multi-robot path planning based on an initiating algorithm. The method is realized by the steps that firstly, the distance between each point is calculated according to the coordinates of a task target point. Then, according to the transition probabilities, the system calculates a transition probability matrix. And finally, according to the transition probability matrix, the system reselects the maximum probability value and the next task target point, calculates the total length of the path, and updates the pheromone concentration between the task target points by using the total length value in a positive feedback mode. The method has the following defects: when the ant colony algorithm generates an initial path scheme, because the environment is not sufficiently perceived and the pheromone is slowly updated, a potential risk of low convergence speed exists in the process of searching the global optimal point. Moreover, due to the characteristic of positive feedback when the pheromone concentration is updated, if the initially obtained solution is a suboptimal solution, the suboptimal solution can quickly occupy the advantages due to the positive feedback, so that the algorithm is trapped in local optimization and is difficult to jump out of the local optimization.
Disclosure of Invention
The invention aims to provide a robot cluster task optimal path planning method based on a fusion algorithm aiming at overcoming the defects of the prior art, and aims to overcome the defects that the pheromone is slow to update due to random path selection and the convergence speed is slow finally caused by the fact that pheromone concentration is repeatedly overlapped due to the positive feedback characteristic, the algorithm is difficult to jump out after being trapped in local optimization, and the solving speed is slow due to the fact that the time cost in the iteration direction of calculation is high.
The idea for realizing the purpose of the invention is as follows: when an initial path is generated, the task target points are coded according to the distribution condition of the task target points, then the robot is rapidly screened and optimized according to the condition of the robot for selecting the path, and the problem of slow convergence rate finally caused by the fact that the original ant colony algorithm needs to randomly select the path in the initial stage is solved. According to the method, when each iteration is finished, the scheme to be selected is effectively expanded through searching of a plurality of paths and a single path, the path scheme of the optimal robot is obtained through the screening operation of the scheme to be selected, and the problem that the optimal solution is obtained in an original ant colony algorithm, the optimal solution finally falls into local optimization due to the positive feedback characteristic, and then the optimal solution is difficult to jump out is solved. After the local search is completed, the current pheromone concentration is obtained by storing and updating the current path scheme and the global path scheme, and the global pheromone concentration is fused to obtain more comprehensive pheromone concentration, so that the relation between the local and the global is balanced, the next iteration direction is determined, and the problem of slow solving speed caused by high time cost of the algorithm for calculating the iteration direction is solved.
The method comprises the following specific steps:
step 1.1, carrying out binary coding on each task target point in each task target point of the task target point cluster by using a binary coding method; connecting the coded task target points into a total path code of the task target points;
step 1.2, generating an initial task target point of each robot in the robot cluster by using a random function, changing the state of the robot after the initial task target point is determined into a planned robot, and splitting a sub-path code of each robot from a total path code;
step 1.3, calculating the sub-path fitness of each planned robot;
step 1.4, sequentially carrying out screening, crossing and mutation genetic operations on each planned robot, and changing the state of the unplanned robot after the genetic operations into the planned robot;
step 1.5, setting the sub-path distance cost of the 0.6M planned robots sequenced in the screening process as a distance cost threshold value D;
step 1.6, decoding the binary coding of the sub-path of each planned robot by utilizing a decimal decoding mode to obtain the sub-path of the robot; calculating the sub-path distance cost of each planned robot, and determining the state of the current planned robot;
step 1.7, judging whether the states of all robots in the robot cluster are planned robots, if so, collecting the sub-path binary codes of all robots as an initial robot cluster path scheme, and continuing to execute the step 2; otherwise, executing step 1.3;
step 2.1, according to the path scheme of the current robot cluster, respectively placing each robot to be updated on an initial task target point corresponding to the robot, and changing the state of the point into visited;
step 2.2, judging whether each placed robot has an unaccessed task target point in the adjacent task target points, if so, executing step 2.3; otherwise, executing step 2.4;
step 2.3, updating a task target point of a next planned path of the robot by using a state transfer main formula, and changing the state of the updated task target point into an accessed state;
step 2.4, updating a task target point of the next planned path of the robot by using a state transfer secondary formula, and changing the state of the updated task target point into an accessed state;
step 2.5, judging whether the states of all task target points in the task target point cluster are visited or not, if yes, executing step 3; otherwise, executing step 2.2;
and 3, updating the task target point of each robot by using a local search method:
step 3.1, decoding the binary coding of the sub-path of each robot by utilizing a decimal decoding method to generate the sub-path corresponding to the robot;
3.2, generating a transfer serial number for the sub-path sequence of each robot by using a random function; according to the generated transfer serial number, inserting a task target point corresponding to the transfer serial number of each robot into the transfer serial number of the next robot sub-path in the sub-path sequence of each robot;
3.3, generating an exchange sequence number for the sub-path sequences of every two robots by using a random function, and exchanging task target points corresponding to the exchange sequence numbers in the sub-paths of the two robots;
3.4, generating two overturning serial numbers for the sub-path sequence of each robot by using a random function, and overturning the path interval between two task target points corresponding to the overturning serial numbers of the robots;
step 3.5, calculating the distance cost of each sub-path, and summing to obtain the total distance cost;
step 3.6, judging that after 5 continuous updates, the difference value between the total distance cost after each update and the total distance cost after the last update is less than 3% of the total distance cost after the update, if so, updating the binary coding of the sub-path of each robot into the current robot cluster path scheme, and executing step 4; otherwise, executing step 3.2;
and 4, updating the pheromone concentration by using an elite strategy:
step 4.1, judging whether the total distance cost of the current robot cluster path scheme is smaller than that of the current global scheme or not, and if so, taking the current robot cluster path scheme as a new global scheme; otherwise, the global scheme is not changed; the current global scheme is that when t =0, the robot cluster path scheme generated in step 1.7 is taken as the current global scheme, and when t ≠ 0, the global scheme updated in the previous round is taken as the current global scheme;
step 4.2, according to a residual pheromone concentration formula: tau is i ' j (t)＝(1-ρ)τ ij (t-1) calculating the pheromone concentration tau remaining after the pheromone concentration to be updated naturally volatilizes in the space in the path between the task target point i and the task target point j i ' j (t); ρ represents a pheromone volatilization factor; the pheromone concentration to be updated is set to be 2 when t =0, and the final pheromone concentration after the previous round of updating is taken as the pheromone concentration to be updated when t is not equal to 0;
step 4.3, according to the following local pheromone release concentration formula, calculating the local pheromone concentration released by the robot on the path between the path task target point i and the task target point j
Wherein L is p (t) represents the total distance cost, L, of the current robot cluster path solution g (T) represents the total distance cost of the global solution, T p Representing a current robot cluster path scheme;
step 4.4, calculating the global pheromone concentration released by the robot on the path between the path task target point i and the task target point j according to the global pheromone release concentration formula
Wherein L is p (t) represents the total distance cost, L, of the current robot cluster path solution g (T) represents the total distance cost of the global solution, T g Representing a global scheme;
and 4.5, calculating the approach of the robot according to the pheromone concentration updating formulaFinal pheromone concentration tau on the path of task target point i and task target point j ij (t)：
Compared with the prior art, the invention has the following advantages:
firstly, the method generates an initial robot cluster path scheme by using a genetic algorithm according to the task target point and the robot selection path condition, overcomes the defect of slow convergence rate finally caused by randomly selecting a path in the initial stage of the prior art, improves the convergence efficiency of the algorithm, skips slow information updating, directly searches the path globally and can obtain the optimal path of the robot more quickly.
Secondly, the invention updates the task target point of each robot by using local search operation between a single path and a plurality of paths, increases the possibility of finding more solutions in iteration, and overcomes the defect that the algorithm is difficult to jump out after being trapped in local optimum due to repeated overlapping of pheromones by the characteristics of positive feedback in the prior art. The invention can obviously improve the diversity of the ant colony algorithm while keeping the current optimal path scheme, so that the algorithm effectively crosses the local extreme point, and the optimal path can be more comprehensively planned for the robot.
Thirdly, the invention adopts an elite strategy to update the task target point of each robot. By means of an elite strategy, a fusion algorithm carries out balanced emphasis on a local optimal solution and a global optimal solution generated in the current iteration turn, and the defect of low solving speed caused by high time cost of calculating the iteration direction in the prior art is overcome.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of a branch local search operation of the present invention;
FIG. 3 is a schematic diagram of the exchange partial search operation of the present invention;
FIG. 4 is a schematic diagram of the flipped local search operation of the present invention;
Detailed Description
The present invention is described in further detail below with reference to the attached drawing figures.
The implementation steps of the present invention are described in further detail with reference to fig. 1.
Step 1.1, performing binary coding on each task target point in each task target point of the task target point cluster by using a binary coding method; and connecting the coded task target points into a total path code of the task target points of the robot cluster. In the embodiment of the present invention, the number of task target points p =50.
Step 1.2, generating an initial task target point of each robot in the robot cluster by using a random function, changing the state of the robot after the initial task target point is determined into a planned robot, taking a binary code of each initial task target point as a splitting starting point of the planned robot placed at the point, taking a binary code of a task target point which is next to the splitting starting point as a splitting end point of the robot, and splitting a sub-path code of each robot from a total path code.
In the embodiment of the invention, there are five robots, which are respectively: robot a, robot B, robot C, robot D, robot E. The initial task target points of the five robots are generated using a random function: the initial task target point of the robot a is the task target point No. 9, the initial task target point of the robot B is the task target point No. 12, the initial task target point of the robot C is the task target point No. 24, the initial task target point of the robot D is the task target point No. 37, and the initial task target point of the robot E is the task target point No. 41. Five robots in a robot cluster: the states of the robot A, the robot B, the robot C, the robot D and the robot E are changed into the planned robots. And taking the binary code of each initial task target point as a splitting starting point of the planned robot placed at the point, and taking the binary code of the task target point which is next to the traveling starting point as a splitting end point of the robot, so that the total path code containing all the task target points is split into 5 sub-path codes corresponding to 5 planned robots. At this time, each planned robot obtains a sub-path code formed by connecting a plurality of task target points, and the sub-path code represents the sequence of the task target points in the future traveling sub-path of the robot.
In an embodiment of the present invention, the total number of robots in a cluster is M =5, and the number of sub-paths M' = M.
Step 1.3, calculating the sub-path fitness of each planned robot:
wherein p is m The sub-path of the mth planned robot is shown, i and j respectively show the task target point of the number i and the task target point of the number j,
representing the path required by the m-th robot path connecting task target point i and task target point jThe distance spent cost.
And 1.4, sequentially carrying out screening, crossing and mutation genetic operations on each planned robot, and carrying out state change on unplanned robots generated in the genetic operations.
And sequencing the fitness of all the planned robots in a high-to-low sequence, reserving the first 0.6M planned robots, and selecting the rest planned robots to be eliminated as random robots.
And sequencing the screened planned robots by using a random function, pairing every two planned robots into a group from front to back, taking the sub-path codes of the paired planned robots as parents, crossing task target points in the two parent path codes, and generating crossed idle sub-path binary codes. The total number of the idle sub-path binary codes after crossing is 0.3M.
And performing negation operation on the task target point in the binary coding of the sub-path of each planned robot according to the probability of 0.2 to obtain the mutated binary coding of the idle sub-path. The total number of the mutated idle sub-path binary codes is 0.6M
Respectively decoding all crossed and mutated idle sub-paths by using a decimal decoding method to obtain 0.9M idle sub-paths, and calculating the distance cost of each idle sub-path; and sequencing all distance costs from small to large, setting the binary coding sequence of the first 0.4M idle sub-paths in the sequencing as the binary coding of the sub-paths of the unplanned robot, and simultaneously changing the state of the unplanned robot into the planned robot.
In the embodiment of the invention, the total number of the crossed idle sub-path binary codes is 2, the total number of the different idle sub-path binary codes is 3, and the total number of the idle sub-paths is 5.
And step 1.5, setting the sub-path distance cost of the 0.6M planned robots sequenced in the screening process as a distance cost threshold value D.
Step 1.6, decoding binary codes of all planned subpaths of the robot by utilizing a decimal decoding mode to obtain subpaths of the robot, summing distances between task target points on the subpaths, and calculating the distance cost of each subpath; summing the distances between the task target points of all paths according to the sequence of the task target points of the paths on each sub-path to obtain the distance cost of the sub-path; and (4) reserving the planned robots with the sub-path distance costs smaller than the distance cost threshold value D, and changing the states of the planned robots with the sub-path costs larger than the distance cost threshold value D into the unplanned robots.
Step 1.7, judging whether the states of all robots in the robot cluster are planned robots, if so, collecting binary codes of sub-paths of all robots as a robot cluster path scheme, and executing step 2; otherwise, step 1.3 is performed.
And 2, updating the task target point of each robot by using a transfer formula.
And 2.1, according to the initial path scheme of the robot cluster, respectively placing all the robots to be updated on initial task target points of the robots, and changing the state of the points into visited ones.
In an embodiment of the present invention, a sub-path binary code of each robot to be updated is obtained from the robot cluster path scheme. Decoding the sub-path binary code of the robot to be updated by utilizing a decimal decoding mode to obtain an updated initial task target point: the initial task target point of the robot a is updated from the task target point No. 9 to the task target point No. 3, the initial task target point of the robot B is updated from the task target point No. 12 to the task target point No. 15, the initial task target point of the robot C is the task target point No. 27 from the task target point No. 24, the initial task target point of the robot D is updated from the task target point No. 37 to the task target point No. 32, and the initial task target point of the robot E is updated from the task target point No. 41 to the task target point No. 49. And respectively placing each robot to be updated on the initial task target point of the robot.
Step 2.2, judging whether each placed robot has an unaccessed task target point in the adjacent task target points, if so, executing step 2.3; otherwise, executing step 2.4;
and 2.3, updating a task target point of the next planned path of the robot by using a state transition main formula. The state transition main formula is as follows.
Wherein,
represents the probability of the mth robot to be updated from the task target point i to the task target point j in the t-th updating, and tau ij (t) indicates the pheromone concentration on the path between task target point i and task target point j, η ij Is a heuristic factor representing the distance d between the task target point i and the task target point j ij The inverse number of (c) is,
s represents a task target node set which is not visited yet, and alpha represents the cooperation degree of the robot; beta indicates the relative importance of pheromone concentration.
And 2.4, updating a task target point of the next planned path of the robot by using a state transfer secondary formula. The state transition sub-formula is as follows.
Wherein j is m And representing the task target point updated by the mth robot to be updated. q is a value selected using a uniform probability random function in the range of [0,1); q. q.s 0 Is a probability threshold. By using the formula, the robot has a probability q 0 Task target node for computing "best sum" and has 1-q 0 The task target node is calculated by using the formula again.
Step 2.5, judging whether the states of all task target points in the task target point cluster are visited, if so, executing step 3; otherwise, executing step 2.2;
in the embodiment of the invention, τ is set when t =0 updates ij (0)＝2，α＝1,β＝3， q 0 ＝0.7。
And 3, updating the task target point of each robot by using local search.
Step 3.1, decoding the sub-path binary code of each robot by utilizing a decimal decoding method to obtain a sub-path corresponding to the robot;
and 3.2, generating a transfer serial number for the sub-path sequence of each robot by using a random function, and inserting a task target point corresponding to the transfer serial number in the sub-path of the robot into the transfer serial number of the next sub-path of the robot.
This transfer step is described in further detail with reference to fig. 2. The sequence of boxes represents the sub-path of each robot, the letters in the front of the sequence of boxes represent the robot number, and the numbers in the sequence of boxes represent the task target point number for that sub-path.
In the embodiment of the present invention, the transfer sequence number generated by the random function for the sub-path sequence of each robot is as follows: the transfer sequence number of the robot a sub-path sequence is 2, the transfer sequence number of the robot B sub-path sequence is 4, the transfer sequence number of the robot C sub-path sequence is 8, the transfer sequence number of the robot D sub-path sequence is 5, and the transfer sequence number of the robot E sub-path sequence is 6.
And according to the generated transfer serial number, inserting the task target point corresponding to the transfer serial number of the robot into the transfer serial number of the next robot sub-path in the sub-path sequence of each robot. Inserting a task target point No. 26 corresponding to the transfer serial number No. 2 into the transfer serial number No. 4 of the sub-path sequence of the robot B in the sub-path sequence of the robot A; inserting a task target point No. 46 corresponding to the transfer serial number No. 4 into the transfer serial number No. 8 of the sub-path sequence of the robot C in the sub-path sequence of the robot B; in the robot C sub-path sequence, inserting a task target point No. 42 corresponding to the transfer serial number No. 8 into the transfer serial number No. 5 of the robot D sub-path sequence; inserting a task target point No. 50 corresponding to the transfer serial number No. 5 into the transfer serial number No. 6 of the sub-path sequence of the robot E in the sub-path sequence of the robot D; and in the robot E sub-path sequence, inserting the task target point No. 33 corresponding to the transfer serial number No. 6 into the transfer serial number No. 2 of the robot A sub-path sequence.
And 3.3, generating an exchange sequence number for the sub-path sequences of every two robots by using a random function, and exchanging task target points corresponding to the exchange sequence numbers in the sub-paths of the two robots.
This exchange step is described in further detail with reference to fig. 3. The box sequence represents the sub-path, the letters in front of the box sequence represent the robot number, and the numbers in the box sequence represent the task target point number of each robot sub-path.
In an embodiment of the invention, the random function generates an exchange sequence number for every two robots: the exchange sequence number between the robot a sub-path and the robot B sub-path is 3, the exchange sequence number between the robot B sub-path and the robot C sub-path is 5, the exchange sequence number between the robot C sub-path and the robot D sub-path is 8, and the exchange sequence number between the robot D sub-path and the robot E sub-path is 2.
And exchanging the task target points corresponding to the exchange sequence numbers in the sub-paths of the two robots according to the generated exchange sequence numbers. Exchanging the task target point No. 1 corresponding to the sub-path sequence No. 3 of the robot A with the task target point No. 20 corresponding to the sub-path sequence No. 3 of the robot B; exchanging the task target point No. 13 corresponding to the sub-path sequence No. 5 of the robot B with the task target point No. 49 corresponding to the sub-path sequence No. 5 of the robot C; exchanging a task target point No. 46 corresponding to the sub-path sequence No. 8 of the robot C with a task target point No. 28 corresponding to the sub-path sequence No. 8 of the robot D; exchanging the task target point No. 31 corresponding to the sub-path sequence No. 2 of the robot D with the task target point No. 12 corresponding to the sub-path sequence No. 2 of the robot E;
and 3.4, generating two overturning serial numbers for the sub-path sequence of each robot by using a random function, and overturning the path interval between two task target points corresponding to the overturning serial numbers of the robots.
This flipping step is described in further detail with reference to fig. 4. The box sequence represents the sub-path, the letters in front of the box sequence represent the robot number, and the numbers in the box sequence represent the task target point number of each robot sub-path.
In an embodiment of the invention, the random function generates two flip sequence numbers for the sub-path sequence of each robot: the turn sequence numbers of the sub-path sequence of robot a are 1 and 6, the turn sequence numbers of the sub-path sequence of robot B are 3 and 9, the turn sequence numbers of the sub-path sequence of robot C are 2 and 4, the turn sequence numbers of the sub-path sequence of robot D are 8 and 10, and the turn sequence numbers of robot E are 5 and 6.
And turning the path interval between the turning serial numbers of each robot according to the generated turning serial numbers. The sub-path section of robot a is flipped over to [41,14,22,20,33,10], the sub-path section of robot B is flipped over to [40,15,45,35,49,26], the sub-path section of robot C is flipped over to [13,27,7,47], the sub-path section of robot D is flipped over to [43,16,46], and the sub-path section of robot E is flipped over to [50,6].
And 3.5, calculating the distance cost of each sub-path, and summing to obtain the total distance cost.
Step 3.6, judging that after 5 continuous updates, the difference value between the total distance cost after each update and the total distance cost after the last update is less than 3% of the total distance cost after the update, if so, updating the binary coding of the sub-path of each robot into the current robot cluster path scheme, and executing step 4; otherwise, step 3.2 is performed.
And 4, updating the pheromone concentration by using an elite strategy.
Step 4.1, judging whether the total distance cost of the current robot cluster path scheme is smaller than that of the current global scheme or not, and if so, taking the current robot cluster path scheme as a new global scheme; otherwise, the global scheme is not changed; the current global solution is to use the robot cluster path solution generated in step 1.7 as the current global solution when t =0, and use the global solution updated in the previous round as the current global solution when t ≠ 0.
Step 4.2, according to a residual pheromone concentration formula: tau' ij (t)＝(1-ρ)τ ij (t-1) calculating a pheromone concentration τ 'remaining after the pheromone concentration to be updated naturally volatilizes in the space in the route between the route task target point i and the task target point j' ij (t); ρ represents a pheromone volatilization factor; the pheromone concentration to be updated is set to be 2 when t =0, and the final pheromone concentration after the previous round of updating is taken as the pheromone concentration to be updated when t ≠ 0.
Step 4.3, according to the following local pheromone release concentration formula, calculating the local pheromone concentration released by the robot on the path between the path task target point i and the task target point j
Wherein L is p (t) represents the total distance cost, L, of the current robot cluster path solution g (T) represents the total distance cost of the global solution, T p Representing a current robot cluster path plan;
step 4.4, calculating the global pheromone concentration released by the robot on the path between the path task target point i and the task target point j according to the global pheromone release concentration formula
Wherein L is p (t) represents the total distance cost, L, of the current robot cluster path solution g (T) represents the total distance cost of the global solution, T g Representing a global scheme.
Step 4.5, according to the pheromone concentration updating formula, calculating the final pheromone concentration tau of the robot on the path between the path task target point i and the task target point j ij (t)。
Claims (10)
1. A robot cluster task optimal path planning method based on a fusion algorithm is characterized in that an initial robot cluster path scheme is generated by a genetic algorithm, a task target point of each robot is updated by local search, and a task target point of each robot is updated by an elite strategy; the method comprises the following steps:
step 1, generating an initial robot cluster path scheme by using a genetic algorithm;
step 1.1, carrying out binary coding on each task target point in each task target point of the task target point cluster by using a binary coding method; connecting the coded task target points into a total path code of the task target points;
step 1.2, generating an initial task target point of each robot in the robot cluster by using a random function, changing the state of the robot after the initial task target point is determined into a planned robot, and splitting a sub-path code of each robot from a total path code;
step 1.3, calculating the sub-path fitness of each planned robot;
step 1.4, sequentially carrying out screening, crossing and mutation genetic operations on each planned robot, and changing the state of the unplanned robot after the genetic operations into the planned robot;
step 1.5, setting the sub-path distance cost of the 0.6M planned robots sequenced in the screening process as a distance cost threshold value D;
step 1.6, decoding the binary codes of the subpaths of each planned robot by utilizing a decimal decoding mode to obtain subpaths of the robot; calculating the sub-path distance cost of each planned robot, and determining the current state of the planned robot;
step 1.7, judging whether the states of all robots in the robot cluster are planned robots, if so, collecting the sub-path binary codes of all robots as an initial robot cluster path scheme, and continuing to execute the step 2; otherwise, executing step 1.3;
step 2, updating the task target point of each robot:
step 2.1, according to the path scheme of the current robot cluster, respectively placing each robot to be updated on an initial task target point corresponding to the robot, and changing the state of the point into visited;
step 2.2, judging whether each placed robot has an unaccessed task target point in the adjacent task target points, if so, executing step 2.3; otherwise, executing step 2.4;
step 2.3, updating a task target point of a next planned path of the robot by using a state transfer main formula, and changing the state of the updated task target point into an accessed state;
step 2.4, updating a task target point of the next planned path of the robot by using a state transfer secondary formula, and changing the state of the updated task target point into an accessed state;
step 2.5, judging whether the states of all task target points in the task target point cluster are visited or not, if yes, executing step 3; otherwise, executing step 2.2;
and 3, updating the task target point of each robot by using a local search method:
step 3.1, decoding the binary coding of the sub-path of each robot by utilizing a decimal decoding method to generate the sub-path corresponding to the robot;
3.2, generating a transfer serial number for the sub-path sequence of each robot by using a random function; according to the generated transfer serial number, inserting a task target point corresponding to the transfer serial number of each robot into the transfer serial number of the next robot sub-path in the sub-path sequence of each robot;
3.3, generating an exchange sequence number for the sub-path sequences of every two robots by using a random function, and exchanging task target points corresponding to the exchange sequence numbers in the sub-paths of the two robots;
3.4, generating two overturning serial numbers for the sub-path sequence of each robot by using a random function, and overturning a path interval between two task target points corresponding to the overturning serial numbers of the robots;
step 3.5, calculating the distance cost of each sub-path, and summing to obtain the total distance cost;
step 3.6, judging that after 5 continuous updates, the difference value between the total distance cost after each update and the total distance cost after the last update is less than 3% of the total distance cost after the update, if so, updating the binary coding of the sub-path of each robot into the current robot cluster path scheme, and executing step 4; otherwise, executing step 3.2;
and 4, updating the pheromone concentration by using an elite strategy:
step 4.1, judging whether the total distance cost of the current robot cluster path scheme is smaller than that of the current global scheme or not, and if so, taking the current robot cluster path scheme as a new global scheme; otherwise, the global scheme is not changed; the current global scheme is that when t =0, the robot cluster path scheme generated in step 1.7 is used as the current global scheme, and when t ≠ 0, the global scheme updated in the previous round is used as the current global scheme;
step 4.2, according to a residual pheromone concentration formula: τ' ij (t)＝(1-ρ)τ ij (t-1) calculating the concentration tau of the pheromone to be updated which is remained after the pheromone concentration naturally volatilizes in the space in the path between the path task target point i and the task target point j i ' j (t); ρ represents a pheromone volatilization factor; the pheromone concentration to be updated is set to be 2 when t =0, and the final pheromone concentration after the previous round of updating is taken as the pheromone concentration to be updated when t ≠ 0;
step 4.3, according to the following local pheromone release concentration formula, calculating the local pheromone concentration released by the robot on the path between the path task target point i and the task target point j
Wherein L is p (t) represents the total distance cost, L, of the current robot cluster path solution g (T) represents the total distance cost of the global solution, T p Representing a current robot cluster path plan;
step 4.4, according to the following global pheromone release concentration formula, calculating the global pheromone concentration released by the robot on the path between the path task target point i and the task target point j
Wherein L is p (t) represents the total distance cost, L, of the current robot cluster path solution g (T) represents the total distance cost of the global solution, T g Represents a global scheme;
step 4.5, according to the pheromone concentration updating formula, calculating the final pheromone concentration tau of the robot on the path between the path task target point i and the task target point j ij (t)：
Step 5, judging whether the updating times reach the optimal fitting updating times, if so, executing step 6; otherwise, executing step 2;
step 6, judging whether the difference value between the total distance cost after each updating and the total distance cost after the last updating is less than 3% of the total distance cost after the iteration updating after 5 times of continuous updating, if so, executing step 7; otherwise, executing step 2;
step 7, decoding the binary coding of the sub-path of each robot by utilizing a decimal decoding method to obtain the sub-path of each robot; and (4) collecting the sub-paths of all the robots as a robot cluster task optimal path planning scheme.
2. The method for planning the optimal path of the robot cluster task based on the fusion algorithm according to claim 1, wherein the specific way of splitting the sub-path code of each robot from the total path code in step 1.2 is as follows: and taking the binary code of each initial task target point as a splitting starting point of the planned robot placed at the point, and taking the binary code of the last task target point of the next splitting starting point as a splitting end point of the robot to obtain the sub-path code of the robot.
3. The method for planning task-optimized paths for robot clusters based on fusion algorithm as claimed in claim 1, wherein the calculating the sub-path fitness of each planned robot in step 1.3 is obtained by the following formula:
wherein p is m Representing the sub-path of the mth planned robot, i and j representing the sub-paths defined by task target point i and task target point j,
and the distance cost of the path connected by the mth robot path task target point i and the task target point j is represented.
4. The method for planning the optimal path of the robot cluster task based on the fusion algorithm according to claim 1, wherein the screening in step 1.4 is performed in a specific manner:
and sequencing the fitness of all the planned robots in a high-to-low sequence, reserving the first 0.6M planned robots, and selecting the rest planned robots to be eliminated as random robots.
5. The method for planning the optimal path of the robot cluster task based on the fusion algorithm according to claim 1, wherein the specific way of intersection in step 1.4 is as follows:
and sequencing the screened planned robots by using a random function, pairing every two planned robots into a group from front to back, taking the sub-path codes of the paired planned robots as parents, crossing task target points in the two parent path codes, and generating crossed idle sub-path binary codes. The total number of the idle sub-path binary codes after crossing is 0.3M.
6. The fusion algorithm-based robot cluster task optimal path planning method according to claim 1, wherein the specific way of variation in step 1.4 is:
and performing negation operation on the task target point in the binary coding of the sub-path of each planned robot according to the probability of 0.2 to obtain the mutated idle sub-path binary coding, wherein the total number of the mutated idle sub-path binary coding is 0.6M.
7. The method for planning the optimal path of the robot cluster task based on the fusion algorithm according to claim 1, wherein the specific way of changing the state of the unplanned robot after genetic manipulation into the planned robot in step 1.4 is:
respectively carrying out decimal decoding on binary codes of all the crossed and mutated idle sub-paths to obtain 0.9M idle sub-paths, and calculating the distance cost of each idle sub-path; and sequencing all distance costs from small to large, setting the binary coding sequence of the first 0.4M idle sub-paths in the sequencing as the binary coding of the sub-paths of the unplanned robot, and simultaneously changing the state of the unplanned robot into the planned robot.
8. The method for planning task optimal path of robot cluster based on fusion algorithm according to claim 1, wherein the step 1.6 of calculating the sub-path distance cost of each planned robot and determining the state of the current planned robot is specifically as follows:
summing the distances between the task target points of all paths according to the sequence of the task target points of the paths on each sub-path to obtain the distance cost of the sub-path; and (4) reserving the planned robots with the sub-path distance costs smaller than the distance cost threshold value D, and changing the states of the planned robots with the sub-path costs larger than the distance cost threshold value D into unplanned robots.
9. The method for planning task-optimized paths of robot clusters based on fusion algorithm according to claim 1, wherein the state transition main formula in step 2.3 is as follows:
wherein,
represents the probability of the mth robot to be updated from the task target point i to the task target point j in the t-th updating, and tau ij (t) the pheromone concentration on the path between task target point i and task target point j, η ij Is a heuristic factor representing the distance d between the task target point i and the task target point j ij The inverse number of (c) is,
s represents a task target node set which is not visited yet, and alpha represents the cooperation degree of the robot; beta indicates the relative importance of pheromone concentration.
10. The fusion algorithm-based robot cluster task optimal path planning method according to claim 1, wherein the state transition subformula in step 2.4 is as follows:
wherein j is m And representing the task target point updated by the mth robot to be updated. q is a value selected using a uniform probability random function in the range of [0,1); q. q.s 0 Is a probability threshold. Using the above formula, the robot has a probability q 0 Task target node for computing "best sum" and has 1-q 0 The probability of the task target node is calculated by using the formula again.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210941660.4A CN115167460A (en) | 2022-08-08 | 2022-08-08 | Robot cluster task optimal path planning method based on fusion algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210941660.4A CN115167460A (en) | 2022-08-08 | 2022-08-08 | Robot cluster task optimal path planning method based on fusion algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115167460A true CN115167460A (en) | 2022-10-11 |
Family
ID=83479432
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210941660.4A Pending CN115167460A (en) | 2022-08-08 | 2022-08-08 | Robot cluster task optimal path planning method based on fusion algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115167460A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116542412A (en) * | 2023-04-28 | 2023-08-04 | 北京大数据先进技术研究院 | Method, device, equipment and medium for processing multitasking operation path conflict |
| CN117114560A (en) * | 2023-10-25 | 2023-11-24 | 中科云谷科技有限公司 | Method, device, storage medium and processor for determining article transport path |
-
2022
- 2022-08-08 CN CN202210941660.4A patent/CN115167460A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116542412A (en) * | 2023-04-28 | 2023-08-04 | 北京大数据先进技术研究院 | Method, device, equipment and medium for processing multitasking operation path conflict |
| CN116542412B (en) * | 2023-04-28 | 2024-02-06 | 北京大数据先进技术研究院 | Method, device, equipment and medium for processing multitasking operation path conflict |
| CN117114560A (en) * | 2023-10-25 | 2023-11-24 | 中科云谷科技有限公司 | Method, device, storage medium and processor for determining article transport path |
| CN117114560B (en) * | 2023-10-25 | 2024-01-05 | 中科云谷科技有限公司 | Method, device, storage medium and processor for determining article transport path |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115167460A (en) | Robot cluster task optimal path planning method based on fusion algorithm | |
| CN110632907B (en) | Scheduling optimization method and system for distributed assembly type replacement flow shop | |
| CN110264062B (en) | Distributed multi-AGV dynamic task allocation and path planning method and system thereof | |
| CN112001541B (en) | Path optimization method based on improved genetic algorithm | |
| CN111382915B (en) | Flexible job shop scheduling method for co-fusion AGVs | |
| CN110802601B (en) | Robot path planning method based on fruit fly optimization algorithm | |
| CN108413963A (en) | Bar-type machine people's paths planning method based on self study ant group algorithm | |
| CN113359702B (en) | Intelligent warehouse AGV operation optimization scheduling method based on water wave optimization-tabu search | |
| CN117314078B (en) | Deadlock-free scheduling method of flexible manufacturing system based on Petri network and neural network | |
| CN117669992B (en) | Intelligent storage multi-mobile robot-oriented real-time two-stage scheduling method and system | |
| CN117852825B (en) | Deadlock-free scheduling method of flexible manufacturing system containing central resources based on deep learning | |
| Kurniasih et al. | Heuristics and metaheuristics approach for query optimization using genetics and memetics algorithm | |
| Kumar et al. | Critical analysis of genetic algorithm under crossover and mutation rate | |
| Kumar et al. | Proposed selection technique of evolutionary algorithm and its implementation for combinatorial problems | |
| Lim et al. | Multi-depot split-delivery vehicle routing problem | |
| Ribeiro et al. | Ant colony optimization algorithm and artificial immune system applied to a robot route | |
| CN113219994A (en) | Grid map optimal path planning method comprising traffic signal lamps | |
| CN112238456A (en) | Material sheet sorting path planning method based on ant colony algorithm | |
| Phuang | The flower pollination algorithm with disparity count process for scheduling problem | |
| Oonsrikaw et al. | Modified ant colony optimization algorithm for multiple-vehicle traveling salesman problems | |
| CN113741482B (en) | Multi-agent path planning method based on asynchronous genetic algorithm | |
| CN113792494B (en) | Multi-target flexible job shop scheduling method based on migration bird swarm algorithm and cross fusion | |
| CN113139774B (en) | Multi-warehouse transportation-oriented vehicle path optimization method | |
| CN115296717A (en) | Scheduling method of ground station antenna array of low-earth-orbit satellite system | |
| CN115700647A (en) | Workshop flexible operation scheduling method based on tabu search genetic algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |