Multi-robot task allocation method based on improved multi-population genetic algorithm

Abstract

The present invention discloses a multi-robot task allocation method based on an improved multi-population genetic algorithm. First, virtual coding is used as the encoding method of chromosomes, and initial feasible chromosomes are generated based on topological sorting. Then, the fitness of chromosomes and populations as a whole is improved through local genetic operators and global genetic operators, that is, the quality of multi-robot task allocation results is improved. Finally, an adaptive termination strategy is introduced to reduce the calculation time of the algorithm. The present invention can minimize the total time required for the entire robot team to visit all target locations while satisfying each timing constraint, and effectively solves the multi-robot task allocation problem under complex timing constraints.

Description

Technical Field

[0001] The invention belongs to the technical field of robot task allocation, and in particular relates to the design of a multi-robot task allocation method based on an improved multi-population genetic algorithm. Background Art

[0002] Multi-robot task allocation refers to effectively allocating multiple tasks to multiple robots based on robot characteristics and system constraints to achieve efficient collaboration of robots. Through centralized or distributed task allocation strategies, robot teams can be scheduled to quickly respond to task requirements and environmental changes, thereby improving production efficiency and enhancing performance. However, with the continuous improvement of social intelligence, multi-robot task allocation faces more complex task requirements, among which timing constraints are prevalent in many industrial and social service scenarios, which are used to outline the priority / order of task completion. For example, in key areas such as logistics distribution and disaster relief, some tasks must be performed before other tasks due to their extremely high importance or urgency. Faced with such challenges, robots not only need to visit multiple target points in sequence, but also must strictly follow the timing constraints between these target points. The core research focus of this problem is how to optimize the robot's task allocation, that is, to determine which target locations each robot should visit so as to minimize the total time required for the entire robot team to visit all target locations while strictly following the timing constraints.

[0003] In the prior art, algorithms for multi-robot task planning problems can be divided into three categories: algorithms based on mixed integer programming models, algorithms based on market trading mechanisms, and intelligent optimization algorithms based on bionic theory. For task planning problems with timing constraints, algorithms based on mixed integer programming models include branch and bound methods: the basic idea is to search all feasible solution spaces for the problem, but with the increase in the number of robots and tasks, the complexity of the solution will increase exponentially, and it is difficult to apply to the rapid task allocation scenario of large-scale robot systems. Algorithms based on market trading mechanisms include iterative auction algorithms: the idea is to gradually approach the optimal auction results through multiple iterations of optimizing bidding and allocation strategies, but the applicable task types are relatively simple. When faced with robot task planning problems with complex timing constraints, this algorithm is often not competent. Intelligent optimization algorithms based on bionic theory include ant colony algorithms: the principle is simple and the applicability is wide. It can obtain suboptimal solutions to large-scale optimization problems in polynomial time, but it is difficult to follow the timing constraints during the execution of the algorithm, and it is necessary to combine corresponding strategies to avoid the generation of infeasible solutions. At the same time, the ant colony algorithm has insufficient local search capabilities, slow convergence speed, and long time consumption. Summary of the invention

[0004] The purpose of the present invention is to propose a multi-robot task allocation method based on an improved multi-population genetic algorithm, which minimizes the total time required for the entire robot team to visit all target locations while satisfying each timing constraint.

[0005] The technical solution of the present invention is: a multi-robot task allocation method based on an improved multi-population genetic algorithm, comprising the following steps:

[0006] S1. Initialize and set the relevant parameters of the robot and the improved multi-population genetic algorithm, and use virtual coding as the encoding method of chromosomes in the improved multi-population genetic algorithm.

[0007] S2. Initialize the population in the improved multi-population genetic algorithm and divide it into multiple sub-populations. Use topological sorting to construct the robot path of each sub-population as the initial chromosome, and mark the elite chromosome by calculating the fitness of each chromosome.

[0008] S3. Use local genetic operators to improve the fitness of chromosomes in each sub-population.

[0009] S4. Based on elite chromosomes, global genetic operators are used to improve the overall fitness of the population.

[0010] S5. Based on the number of changes in the elite chromosome, an adaptive termination strategy is used to output the multi-robot task allocation results.

[0011] Furthermore, step S1 includes the following sub-steps:

[0012] S11. Initialize and set the robot's related parameters, including the number of target positions. , target location set 、Number of robots , Robot Collection , robot initial position set , the speed of the robot and the robot at the target location and target location Time required to move between ,in , represents the set of target position pairs, represents each pair of target locations with strong timing constraints, Represents each pair of target locations with weak timing constraints, which are represented by a directed graph express, A set of vertices representing the target location, Represents a set of directed edges.

[0013] S12. Initialize and set the parameters related to the improved multi-population genetic algorithm, including the maximum number of evolution generations. , Initial maximum number of evolutionary stagnation , crossover probability , mutation probability , initial iteration count , Initial Stagnation Count and the number of changes in the initial elite chromosomes .

[0014] S13, using virtual coding as the encoding method of chromosomes in the improved multi-population genetic algorithm, using the set express genes at the target location and Insertion into the gene at the target location marker genes, which are composed of express.

[0015] Further, step S2 includes the following sub-steps:

[0016] S21. Initialize the population in the improved multi-population genetic algorithm and divide it into subpopulations, each of which includes Chromosome individuals.

[0017] S22, using the topological sorting method, each time from the directed graph A target position with an in-degree of zero is randomly selected in the graph and the gene represented by it is inserted into the chromosome in sequence. Delete the vertices and directed edges corresponding to the target position with zero in-degree.

[0018] S23, in response to all genes represented by the target positions with in-degree zero being inserted into the chromosome, randomly insert marker genes.

[0019] S24. Calculate the fitness of each chromosome:

[0020]

[0021] in represents the objective function, Indicates In the subpopulation The fitness of a chromosome, , .

[0022] S25, the chromosome with the highest marker fitness is the elite chromosome .

[0023] Furthermore, the local genetic operator in step S3 includes a selection operator, a crossover operator and a mutation operator.

[0024] The selection operator adopts a binary tournament selection algorithm, randomly selects two chromosomes from the current subpopulation, compares their fitness values, and selects the chromosome with higher fitness to enter the next generation, repeating the iteration until the new subpopulation size is the same as the original subpopulation size.

[0025] The crossover operator is specifically: in response to a randomly generated value between 0 and 1 being less than the crossover probability , a hierarchical sorting strategy is used to classify the target position, and according to the classification level, the crossover operator I is used to perform a crossover operation on two different parent chromosomes, and the crossover operator II is used to perform a crossover operation on two identical parent chromosomes.

[0026] The mutation operator is as follows: In response to a randomly generated value between 0 and 1 being less than the mutation probability , the two variant positions on the chromosome and Make an exchange.

[0027] Further, the crossover operator I comprises the following steps:

[0028] A1. Randomly select two parent chromosomes that have not undergone crossover operations.

[0029] A2. Randomly select a target gene on the parent chromosome , and exchange All genes from then to before the marker gene.

[0030] A3. Let the missing gene set be , The gene with the lowest rank in the target position is the chromosome that has undergone the gene exchange operation in step A2. In, confirm Insertion position :

[0031]

[0032] in Represents chromosome The position on , Represents chromosome Length, Indicates that Insertion into chromosome Location The previous objective function value, Indicates that Insertion into chromosome Location The objective function value after that.

[0033] A4. Insertion into chromosome Location , and update the missing gene set and chromosome:

[0034]

[0035]

[0036] in represents the updated missing gene set, Indicates a delete operation. represents the updated chromosome, Indicates that Insertion into chromosome Location .

[0037] Furthermore, the crossover operator II includes the following steps:

[0038] B1. Obtain a gene with the same target position level in one of the parent chromosomes and .

[0039] B2. Gene and First, cut gene fragments of equal length and make sure that the cut fragments do not contain the marker gene.

[0040] B3. Select two gene segments corresponding to those in step B2 in another parent chromosome and perform a crossover operation in an X shape.

[0041] Furthermore, the global genetic operator in step S4 includes a crossover mechanism across populations, deletion of chromosomes that violate timing constraints in each subpopulation, and an elite sharing strategy.

[0042] The cross-population crossover mechanism is as follows: randomly select two sub-populations that have not undergone cross-population crossover and Perform crossover between subpopulations, that is, Randomly selected from Chromosomes and subpopulations Randomly selected from The chromosomes are crossed and the offspring with the highest fitness that does not violate the timing constraints are selected. subpopulations and Among them , ,and .

[0043] The elite sharing strategy is specifically: using elite chromosomes Replace the chromosome with the lowest fitness in each subpopulation.

[0044] Further, step S5 includes the following sub-steps:

[0045] S51. Assume that the chromosome with the highest current fitness value is , determine whether If so, then , and proceed to step S52, otherwise proceed to step S53, where Represents the fitness function.

[0046] S52, using adaptive termination strategy to update the maximum number of evolutionary stagnation times and update the stagnation count , and the number of changes in the elite chromosomes Add 1:

[0047]

[0048] in Indicates elite chromosome occurrence The maximum number of evolutionary stagnation after changes, Indicates elite chromosome occurrence The maximum number of evolutionary stagnation after changes, is the decreasing factor.

[0049] S53, set iteration count Add 1, stop counting Add 1.

[0050] S54: Determine whether and If yes, return to step S3, otherwise output the elite chromosome As a result of multi-robot task allocation.

[0051] The beneficial effects of the present invention are as follows: based on the traditional multi-population genetic algorithm, the present invention effectively improves the fitness of chromosomes and the entire population through local genetic operators and global genetic operators, thereby improving the quality of multi-robot task allocation results, and at the same time introduces an adaptive termination strategy to reduce the calculation time of the algorithm. It can minimize the total time required for the entire robot team to visit all target locations while meeting each timing constraint, effectively solving the multi-robot task allocation problem under complex timing constraints. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 Shown is a flow chart of a multi-robot task allocation method based on an improved multi-population genetic algorithm provided by an embodiment of the present invention.

[0053] Figure 2 Shown is a schematic diagram of virtual encoding provided by an embodiment of the present invention.

[0054] Figure 3 FIG. 4 is a schematic diagram of a hierarchical sorting strategy provided by an embodiment of the present invention.

[0055] Figure 4 FIG. 4 is a schematic diagram of a crossover operator I provided in an embodiment of the present invention.

[0056] Figure 5 FIG. 4 is a schematic diagram of a crossover operator II provided in an embodiment of the present invention.

[0057] Figure 6 FIG. 4 is a schematic diagram of a mutation operator provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] Now, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be understood that the embodiments shown and described in the accompanying drawings are only exemplary and are intended to explain the principles and spirit of the present invention, rather than to limit the scope of the present invention.

[0059] The embodiment of the present invention provides a multi-robot task allocation method based on an improved multi-population genetic algorithm, such as Figure 1 As shown, the following steps S1 to S5 are included:

[0060] S1. Initialize and set the relevant parameters of the robot and the improved multi-population genetic algorithm, and use virtual coding as the encoding method of chromosomes in the improved multi-population genetic algorithm.

[0061] Step S1 includes the following sub-steps S11 to S13:

[0062] S11. Initialize and set the robot's related parameters, including the number of target positions. , target location set 、Number of robots , Robot Collection , robot initial position set , the speed of the robot and the robot at the target location and target location Time required to move between ,in , represents the set of target position pairs, represents each pair of target locations with strong timing constraints, Represents each pair of target locations with weak timing constraints, which are represented by a directed graph express, A set of vertices representing the target location, In the embodiment of the present invention, the strong timing constraint between two target positions means that the same robot should visit these two positions without interruption, and the weak timing constraint between two target positions means that the time of visiting one target position should be earlier or later than that of visiting another target position.

[0063] S12. Initialize and set the parameters related to the improved multi-population genetic algorithm, including the maximum number of evolution generations. , Initial maximum number of evolutionary stagnation , crossover probability , mutation probability , initial iteration count , Initial Stagnation Count and the number of changes in the initial elite chromosomes .

[0064] S13, using virtual coding as the encoding method of chromosomes in the improved multi-population genetic algorithm, using the set express genes at the target location and Insertion into the gene at the target location marker genes, which are composed of express.

[0065] like Figure 2 Shown is a chromosome structure containing 12 target sites and 3 robots, with the target sites divided into three segments by marker genes 13 and 14.

[0066] S2. Initialize the population in the improved multi-population genetic algorithm and divide it into multiple sub-populations. Use topological sorting to construct the robot path of each sub-population as the initial chromosome, and mark the elite chromosome by calculating the fitness of each chromosome.

[0067] Step S2 includes the following sub-steps S21 to S25:

[0068] S21. Initialize the population in the improved multi-population genetic algorithm and divide it into subpopulations, each of which includes Chromosome individuals.

[0069] S22, using the topological sorting method, each time from the directed graph A target position with an in-degree of zero is randomly selected in the graph and the gene represented by it is inserted into the chromosome in sequence. Delete the vertices and directed edges corresponding to the target position with zero in-degree.

[0070] S23, in response to all genes represented by the target positions with in-degree zero being inserted into the chromosome, randomly insert marker genes.

[0071] S24. Calculate the fitness of each chromosome:

[0072]

[0073] in represents the objective function, Indicates In the subpopulation The fitness of a chromosome, , .

[0074] S25, the chromosome with the highest marker fitness is the elite chromosome .

[0075] S3. Use local genetic operators to improve the fitness of chromosomes in each sub-population.

[0076] In the embodiment of the present invention, the local genetic operator includes a selection operator, a crossover operator and a mutation operator.

[0077] Among them, the selection operator adopts the binary tournament selection algorithm, randomly selects two chromosomes from the current subpopulation, compares their fitness values, and selects the chromosome with higher fitness to enter the next generation, repeating the iteration until the new subpopulation size is the same as the original subpopulation size.

[0078] The crossover operator is specifically: in response to a randomly generated value between 0 and 1 being less than the crossover probability , a hierarchical sorting strategy is used to classify the target position, and according to the classification level, the crossover operator I is used to perform a crossover operation on two different parent chromosomes, and the crossover operator II is used to perform a crossover operation on two identical parent chromosomes.

[0079] In the embodiment of the present invention, the hierarchical sorting strategy is based on Figure 3 Taking (a) as an example, the target positions with zero in-degree are 1 and 6, which are classified as the first level. At the same time, the target positions 1 and 6 and the edges from these vertices will be deleted. At this time, the target positions with zero in-degree become 2 and 3, which are classified as the second level. Similarly, these target positions and their corresponding edges will also be deleted. Repeat this process until all target positions are classified, such as Figure 3 (b), where Level represents the level. In addition, for the target position , only when the target position and Only when their in-degrees are all zero, they are allowed to be graded, and their grades are the same.

[0080] like Figure 4 As shown, the crossover operator I includes the following steps A1~A4:

[0081] A1. Randomly select two parent chromosomes that have not undergone crossover operation.

[0082] A2. Randomly select a target gene on the parent chromosome , and exchange All genes from then to before the marker gene.

[0083] In the embodiment of the present invention, some genes need to be deleted from the chromosome before receiving the exchanged genes to ensure that each target position gene can only appear once on the chromosome.

[0084] A3. Let the missing gene set be , The gene with the lowest rank in the target position is the chromosome that has undergone the gene exchange operation in step A2. In, confirm Insertion position :

[0085]

[0086] in Represents chromosome The position on , Represents chromosome Length, Indicates that Insertion into chromosome Location The previous objective function value, Indicates that Insertion into chromosome Location The objective function value after that.

[0087] A4. Insertion into chromosome Location , and update the missing gene set and chromosome:

[0088]

[0089]

[0090] in represents the updated missing gene set, Indicates a delete operation. represents the updated chromosome, Indicates that Insertion into chromosome Location .

[0091] like Figure 5 As shown, the crossover operator II includes the following steps B1 to B3:

[0092] B1. Obtain a gene with the same target position level in one of the parent chromosomes and .

[0093] B2. Gene and First, cut gene fragments of equal length and make sure that the cut fragments do not contain the marker gene.

[0094] B3. Select two gene segments corresponding to those in step B2 in another parent chromosome and perform a crossover operation in an X shape.

[0095] like Figure 6 As shown, the mutation operator is specifically: in response to the randomly generated value between 0 and 1 being less than the mutation probability , the two variant positions on the chromosome and Make an exchange.

[0096] S4. Based on elite chromosomes, global genetic operators are used to improve the overall fitness of the population.

[0097] In the embodiment of the present invention, the global genetic operator includes a crossover mechanism across populations, deletion of chromosomes that violate timing constraints in each subpopulation, and an elite sharing strategy.

[0098] The cross-population crossover mechanism is as follows: randomly select two sub-populations that have not undergone cross-population crossover. and Perform crossover between subpopulations, that is, Randomly selected from Chromosomes and subpopulations Randomly selected from The chromosomes are crossed and the offspring with the highest fitness that does not violate the timing constraints are selected. subpopulations and Among them , ,and .

[0099] In the embodiment of the present invention, the cross-population crossover mechanism allows chromosomes between different populations to exchange genes, which helps to search in a wider solution space, thereby increasing the possibility of finding the global optimal solution. Compared with the crossover within a single population, the cross-population crossover can introduce more diversity, avoid the algorithm from falling into the local optimum too early, and enhance the global search capability. At the same time, through cross-population crossover, the evolutionary process between different populations is no longer isolated, but interconnected and mutually reinforcing. This co-evolution helps various populations share useful information in the evolutionary process and accelerates the optimization process of the entire algorithm.

[0100] The elite sharing strategy is specifically: using elite chromosomes Replace the chromosome with the lowest fitness in each subpopulation.

[0101] In the embodiment of the present invention, the elite sharing strategy periodically selects the best chromosomes from various populations to form an elite population, and retains these chromosomes from participating in subsequent genetic operations (such as crossover and mutation), thereby ensuring that the discovered optimal solution will not be lost during the evolution process. At the same time, elite chromosomes usually have higher fitness values, and their performance in genetic operations is often better than that of ordinary chromosomes. Therefore, through the elite sharing strategy, these excellent chromosomes can spread their excellent genes among different populations, accelerate the evolution speed of the entire population, and thus accelerate the convergence of the algorithm.

[0102] S5. Based on the number of changes in the elite chromosome, an adaptive termination strategy is used to output the multi-robot task allocation results.

[0103] Step S5 includes the following sub-steps S51 to S54:

[0104] S51. Assume that the chromosome with the highest current fitness value is , determine whether If so, then , and proceed to step S52, otherwise proceed to step S53, where Represents the fitness function.

[0105] S52, using adaptive termination strategy to update the maximum number of evolutionary stagnation times and update the stagnation count , and the number of changes in the elite chromosomes Add 1:

[0106]

[0107] in Indicates elite chromosome occurrence The maximum number of evolutionary stagnation after changes, Indicates elite chromosome occurrence The maximum number of evolutionary stagnation after changes, is the decreasing factor. The adaptive termination strategy can effectively avoid wasting too much time due to the difficulty in finding a better solution in subsequent iterations.

[0108] S53, set iteration count Add 1, stop counting Add 1.

[0109] S54, determine whether it is satisfied and If yes, return to step S3, otherwise output the elite chromosome As a result of multi-robot task allocation.

[0110] In the embodiment of the present invention, the adaptive termination strategy changes the termination condition of the algorithm from a single maximum number of iterations to a condition jointly determined by the maximum number of iterations and the maximum number of stagnations.

[0111] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A multi-robot task allocation method based on an improved multi-population genetic algorithm, characterized in that: The following steps are involved: S1. Initialize and set the relevant parameters of the robot and the improved multi-population genetic algorithm, and use virtual coding as the encoding method of chromosomes in the improved multi-population genetic algorithm; S2, initialize the population in the improved multi-population genetic algorithm and divide it into multiple sub-populations, use topological sorting to construct the robot path of each sub-population as the initial chromosome, and mark the elite chromosome by calculating the fitness of each chromosome; S3, using local genetic operators to improve the fitness of chromosomes in each sub-population; S4, based on elite chromosomes, using global genetic operators to improve the overall fitness of the population; S5. Based on the number of changes in the elite chromosomes, an adaptive termination strategy is used to output the multi-robot task allocation results; The step S1 comprises the following sub-steps: S11, initializing and setting the relevant parameters of the robot, the relevant parameters of the robot including the number of target positions , target location set 、Number of robots , Robot Collection , robot initial position set , the speed of the robot and the robot at the target location and target location Time required to move between ,in , represents the set of target position pairs, represents each pair of target locations with strong timing constraints, Represents each pair of target locations with weak timing constraints, which are represented by a directed graph express, A set of vertices representing the target location, represents a set of directed edges; S12, initializing and setting the relevant parameters of the improved multi-population genetic algorithm, wherein the relevant parameters of the improved multi-population genetic algorithm include the maximum evolutionary generation , Initial maximum number of evolutionary stagnation , crossover probability , mutation probability , initial iteration count , Initial Stagnation Count and the number of changes in the initial elite chromosomes ; S13, using virtual coding as the encoding method of chromosomes in the improved multi-population genetic algorithm, using the set express genes at the target location and Insertion into the gene at the target location marker genes, the marker genes are composed of a set express.

2. The multi-robot task allocation method based on improved multi-population genetic algorithm according to claim 1 is characterized in that: The step S2 comprises the following sub-steps: S21. Initialize the population in the improved multi-population genetic algorithm and divide it into subpopulations, each of which includes Chromosome individuals; S22, using the topological sorting method, each time from the directed graph A target position with an in-degree of zero is randomly selected in the graph and the gene represented by it is inserted into the chromosome in sequence. Delete the vertices and directed edges corresponding to the target position with zero in-degree; S23, in response to all genes represented by the target positions with in-degree zero being inserted into the chromosome, randomly insert marker genes; S24. Calculate the fitness of each chromosome: ； in represents the objective function, Indicates In the subpopulation The fitness of a chromosome, , ; S25, the chromosome with the highest marker fitness is the elite chromosome .

3. The multi-robot task allocation method based on improved multi-population genetic algorithm according to claim 2 is characterized in that: The local genetic operators in step S3 include a selection operator, a crossover operator and a mutation operator; The selection operator adopts a binary tournament selection algorithm to randomly select two chromosomes from the current sub-population, compare their fitness values, and select the chromosome with higher fitness to enter the next generation, and repeat the iteration until the new sub-population size is the same as the original sub-population size; The crossover operator is specifically: in response to a randomly generated value between 0 and 1 being less than the crossover probability , a hierarchical sorting strategy is used to classify the target position, and according to the classification level, the crossover operator I is used to perform a crossover operation on two different parent chromosomes, and the crossover operator II is used to perform a crossover operation on two identical parent chromosomes; The mutation operator is specifically: in response to a randomly generated value between 0 and 1 being less than the mutation probability , the two variant positions on the chromosome and Make an exchange.

4. The multi-robot task allocation method based on improved multi-population genetic algorithm according to claim 3 is characterized in that: The crossover operator I comprises the following steps: A1. Randomly select two parent chromosomes that have not undergone crossover operations; A2. Randomly select a target gene on the parent chromosome , and exchange All genes from there to the marker gene; A3. Let the missing gene set be , The gene with the lowest rank in the target position is the chromosome that has undergone the gene exchange operation in step A2. In, confirm Insertion position : ； in Represents chromosome The position on , Represents chromosome Length, Indicates that Insertion into chromosome Location The previous objective function value, Indicates that Insertion into chromosome Location The objective function value after that; A4. Insertion into chromosome Location , and update the missing gene set and chromosome: ； ； in represents the updated missing gene set, Indicates a delete operation. represents the updated chromosome, Indicates that Insertion into chromosome Location .

5. The multi-robot task allocation method based on improved multi-population genetic algorithm according to claim 3 is characterized in that: The crossover operator II comprises the following steps: B1. Obtain a gene with the same target position level in one of the parent chromosomes and ; B2. Gene and First, cut the gene fragments of equal length and ensure that the cut fragments do not contain the marker gene; B3. Select two gene segments corresponding to those in step B2 in another parent chromosome and perform a crossover operation in an X shape.

6. The multi-robot task allocation method based on improved multi-population genetic algorithm according to claim 2 is characterized in that: The global genetic operator in step S4 includes a crossover mechanism across populations, deletion of chromosomes that violate timing constraints in each subpopulation, and an elite sharing strategy; The cross-population crossover mechanism is specifically: randomly select two sub-populations that have not undergone cross-population crossover and Perform crossover between subpopulations, that is, Randomly selected from Chromosomes and subpopulations Randomly selected from The chromosomes are crossed and the offspring with the highest fitness that does not violate the timing constraints are selected. subpopulations and Among them , ,and ; The elite sharing strategy is specifically: using elite chromosomes Replace the chromosome with the lowest fitness in each subpopulation.

7. The multi-robot task allocation method based on improved multi-population genetic algorithm according to claim 2 is characterized in that: The step S5 comprises the following sub-steps: S51. Assume that the chromosome with the highest current fitness value is , determine whether If so, then , and proceed to step S52, otherwise proceed to step S53, where represents the fitness function; S52, using adaptive termination strategy to update the maximum number of evolutionary stagnation times and update the stagnation count , and the number of changes in the elite chromosomes Add 1: ； in Indicates elite chromosome occurrence The maximum number of evolutionary stagnation after changes, Indicates elite chromosome occurrence The maximum number of evolutionary stagnation after changes, is the decreasing factor; S53, set iteration count Add 1, stop counting Add 1; S54: Determine whether and If yes, return to step S3, otherwise output the elite chromosome As a result of multi-robot task allocation.

Images