An AGV real-time transportation and allocation method and device

Abstract

The present application provides an AGV real-time transportation allocation method and device. In the present application, the current picking and scheduling strategy of the AGV performs genetic iteration on a randomly generated set of picking and scheduling chromosomes based on the picking efficiency of the pickers. At the same time, the current picking and scheduling strategy of the AGV can be adjusted in real time according to the real-time picking efficiency transformation of the pickers, and the current picking and scheduling strategy of the AGV can be adapted to the real-time picking efficiency change of the pickers. This method determines the optimal personnel scheduling method at the current moment with a fixed number of iterations for the real-time picking efficiency and scheduling method of the pickers, and then dynamically adjusts the current picking and scheduling strategy of the AGV based on this method, so as to avoid the problem that the picking efficiency of the pickers becomes the bottleneck restricting the overall efficiency of the system and improve the overall operation efficiency of the system.

Description

Technical Field

[0001] The present application relates to the technical field of AGV, and in particular, to an AGV real-time transportation allocation method and device. Background Art

[0002] In the intelligent era, the digitization and unmanned operation of warehousing have become inevitable. Along the path of unmanned operation from fully manual operation without machines, to semi-unmanned operation with human-machine collaboration, and then to fully unmanned operation without humans, the progress of different industries will vary. Due to the complexity and diversity of human work, fully manual and fully unmanned work scenarios will be in the minority in the foreseeable future. Therefore, how to improve efficiency in the semi-unmanned scenario of human-machine collaboration will become an issue for all industries.

[0003] In the process of picking and shipping goods in the modern warehousing industry, the mode has changed from the original traditional warehousing's "people looking for goods" mode to the modern intelligent warehousing's "goods looking for people" mode, which is completed by the collaboration of employees and AGV robots. Specifically, as Figure 1 shown in the schematic diagram of the AGV distribution scenario in the "goods looking for people" mode, the AGV robot carries the shelf from the shelf storage area to in front of the picker's picking workstation. After the picker completes the picking, the AGV robot carries the shelf back to the shelf storage area.

[0004] However, in the process of using AGV robots for human-machine collaborative picking of goods, attention is often only paid to improving the transfer efficiency in the picking process by optimizing the route planning of AGV robots, ignoring the efficiency differences of different pickers picking different products, which makes the picking efficiency of pickers become the bottleneck restricting the overall efficiency of the system.

[0005] Based on the current "goods to person" picking method, this invention focuses on the picking efficiency of pickers, designs an AGV transportation allocation strategy adapted to the current picking efficiency of pickers, aiming to cooperate with the picking efficiency of pickers to improve the overall operation efficiency of the system. Summary of the Invention

[0006] The present application aims to solve at least one of the above technical defects. In view of this, the present application provides an AGV real-time transportation allocation method and device, which are used to solve the technical defect that in the prior art, the efficiency differences of different pickers picking different products are ignored, resulting in the picking efficiency of pickers becoming the bottleneck restricting the overall efficiency of the AGV transportation system.

[0007] To achieve the above object, the following solutions are proposed:

[0008] An AGV real-time transportation allocation method includes:

[0009] Step 1: Generate a picking schedule chromosome set based on preset picking situation parameters. There are a preset number of picking schedule chromosomes randomly arranged in the picking schedule chromosome set;

[0010] Step 2: According to the picking efficiency of the current picking employees, use the preset mathematical programming model to decode each picking schedule chromosome in the picking schedule chromosome set, determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set, as well as the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the target picking schedule strategy;

[0011] Step 3: Based on the fitness value of each picking schedule chromosome in the picking schedule chromosome set, calculate the offspring genetic probability of each picking schedule chromosome, and determine two parent picking schedule chromosomes participating in offspring inheritance in the picking schedule chromosome set according to the offspring genetic probability;

[0012] Step 4: Recombine the gene blocks in the two parent picking schedule chromosomes according to the preset crossover probability and mutation probability to generate two offspring picking schedule chromosomes;

[0013] Step 5: Use the preset mathematical programming model to solve the total backlog of the two offspring picking schedule chromosomes, as well as the corresponding fitness value;

[0014] Step 6: Make a replacement determination according to the fitness values of the two parent picking schedule chromosomes, the fitness values of the two offspring picking schedule chromosomes, and the preset annealing parameter, and generate an offspring schedule chromosome set based on the determination result;

[0015] Step 7: Take the preset fixed number of iterations as the iteration termination condition, and iteratively execute Steps 2 to 6. The offspring schedule chromosome set generated each time will participate in the iteration as the picking schedule chromosome set in the next iteration;

[0016] Step 8: After the iteration terminates, use the preset mathematical programming model to decode each picking schedule chromosome in the offspring schedule chromosome set generated in the last iteration, determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set, as well as the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the current picking schedule strategy;

[0017] Step 9: Determine the current AGV transportation allocation strategy based on the current picking schedule strategy.

[0018] Preferably, in the second step, according to the picking efficiency of the current picker, each picking schedule chromosome in the picking schedule chromosome set is decoded by using a preset mathematical programming model to determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, including:

[0019] According to the picking efficiency of the picker, each picking schedule chromosome in the picking schedule chromosome set is decoded by using a preset mathematical programming model to determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set;

[0020] According to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set, the fitness value of each picking schedule chromosome in the picking schedule chromosome set is determined.

[0021] Preferably, the calculation formula for determining the total backlog of each picking schedule chromosome in the picking schedule chromosome set is:

[0022]

[0023]

[0024] where a ijn describes the duration assigned to product type n by each picker i in picking group j, N is the total number of product types, W is the total number of pickers, B n is the total number of products to be sorted in the nth product type, M in is the picking time for picker i to pick the nth product type, which is also the reciprocal of the skill level of picker i for picking the nth product type, RO j is the remaining number of products to be picked in picking group j, J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

[0025] Preferably, in the eighth step, the objective function is:

[0026]

[0027] where RO j is the remaining number of products to be picked in picking group j, J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

[0028] Preferably, according to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set, determining the fitness value of each picking schedule chromosome in the picking schedule chromosome set includes:

[0029] Determine whether the total backlog of each picking schedule chromosome in the picking schedule chromosome set meets the constraint conditions;

[0030] Determine the fitness value of the picking scheduling chromosome that meets the described constraint conditions as the reciprocal of the total backlog of the picking scheduling chromosome;

[0031] Determine the fitness value of the picking scheduling chromosome that does not meet the described constraint conditions as zero.

[0032] Preferably, in step four, recombinating the gene blocks in the two parental picking scheduling chromosomes according to a preset crossover probability and mutation probability to generate two offspring picking scheduling chromosomes, including:

[0033] Perform a crossover determination according to the preset crossover probability. If the result of the crossover determination is passed, randomly select two corresponding crossover points on the two parental picking scheduling chromosomes for gene block crossover transformation; otherwise, keep the two parental picking scheduling chromosomes unchanged. The result generated after the crossover determination is two crossover transformation scheduling chromosomes;

[0034] Perform a mutation determination according to the preset mutation probability. If the result of the mutation determination is passed, randomly select a single gene block in the two crossover transformation scheduling chromosomes for replacement; otherwise, keep the two crossover transformation scheduling chromosomes unchanged. The result generated after the mutation determination is two mutation transformation scheduling chromosomes;

[0035] Take the two mutation transformation scheduling chromosomes as the two offspring picking scheduling chromosomes.

[0036] Preferably, in step three, based on the fitness value of each picking scheduling chromosome in the picking scheduling chromosome set, calculate the offspring genetic probability of each picking scheduling chromosome, and determine two parental picking scheduling chromosomes participating in offspring inheritance in the picking scheduling chromosome set according to the offspring genetic probability, including:

[0037] Based on the fitness value of each picking scheduling chromosome in the picking scheduling chromosome set, calculate the offspring genetic probability of each picking scheduling chromosome. The offspring genetic probability is the proportion of the fitness value of each picking scheduling chromosome in the total sum of the fitness values of all picking scheduling chromosomes in the picking scheduling chromosome set;

[0038] Adopt the roulette wheel strategy, and based on the offspring genetic probability of each picking scheduling chromosome, determine two parental picking scheduling chromosomes participating in offspring inheritance in the picking scheduling chromosome set.

[0039] Preferably, in step six, perform a replacement determination according to the fitness values of the two parental picking scheduling chromosomes, the fitness values of the two offspring picking scheduling chromosomes, and a preset annealing parameter, and generate an offspring scheduling chromosome set based on the determination result, including:

[0040] Determine the annealing probability according to the preset annealing parameters;

[0041] Compare the fitness values of the two parent picking and scheduling chromosomes and the fitness values of the two offspring picking and scheduling chromosomes, and determine the chromosome with the larger fitness value as the better chromosome;

[0042] If the better chromosome is the offspring picking and scheduling chromosome, replace the parent picking and scheduling chromosome in the picking and scheduling chromosome set with the offspring picking and scheduling chromosome, and obtain the offspring scheduling chromosome set;

[0043] If the better chromosome is the parent picking and scheduling chromosome, determine whether to replace it according to the annealing probability. If the determination result is yes, replace the parent picking and scheduling chromosome in the picking and scheduling chromosome set with the offspring picking and scheduling chromosome, and obtain the offspring scheduling chromosome set. Otherwise, keep the parent picking and scheduling chromosome in the picking and scheduling chromosome set unchanged, and use the picking and scheduling chromosome set as the offspring picking and scheduling chromosome.

[0044] Preferably, the calculation formula of the annealing probability is:

[0045]

[0046] where f(x) is the fitness value of the parent picking and scheduling chromosome, f′(x) is the fitness value of the offspring picking and scheduling chromosome, and T is the temperature.

[0047] An AGV real-time transportation allocation device, comprising:

[0048] A chromosome generation unit, configured to generate a picking and scheduling chromosome set based on preset picking situation parameters, where there are a preset number of picking and scheduling chromosomes generated by random permutation in the picking and scheduling chromosome set;

[0049] An iterative unit, which is used to decode each picking schedule chromosome in the picking schedule chromosome set according to the picking efficiency of the current picker by using a preset mathematical programming model, determine the total backlog volume and the corresponding fitness value of each picking schedule chromosome in the picking schedule chromosome set, and determine the current optimal picking schedule chromosome based on the objective function to generate the target picking schedule strategy; calculate the offspring genetic probability of each picking schedule chromosome in the picking schedule chromosome set according to the fitness value of each picking schedule chromosome in the picking schedule chromosome set, and determine two parental picking schedule chromosomes participating in offspring inheritance in the picking schedule chromosome set according to the offspring genetic probability; recombine the gene blocks in the two parental picking schedule chromosomes according to the preset crossover probability and mutation probability to generate two offspring picking schedule chromosomes; use the preset mathematical programming model to solve the total backlog volume and the corresponding fitness value of the two offspring picking schedule chromosomes; perform replacement determination according to the fitness values of the two parental picking schedule chromosomes, the fitness values of the two offspring picking schedule chromosomes, and the preset annealing parameters, and generate an offspring schedule chromosome set based on the determination result; use the preset fixed number of iterations as the iteration termination condition, and iteratively execute steps two to six. The offspring schedule chromosome set generated each time will participate in the iteration as the picking schedule chromosome set in the next iteration;

[0050] An allocation strategy unit, which is used to, after the iteration terminates, decode each picking schedule chromosome in the offspring schedule chromosome set generated in the last iteration by using a preset mathematical programming model, determine the total backlog volume and the corresponding fitness value of each picking schedule chromosome in the picking schedule chromosome set, and determine the current optimal picking schedule chromosome based on the objective function to generate the current picking schedule strategy; determine the current AGV transportation allocation strategy based on the current picking schedule strategy.

[0051] As can be seen from the above technical solution, an AGV real-time delivery allocation method and device provided by an embodiment of the present application generate a picking schedule chromosome set through preset picking situation parameters. According to the picking efficiency of the current pickers, each picking schedule chromosome in the picking schedule chromosome set is decoded by using a preset mathematical programming model to determine the total backlog amount and the corresponding fitness value of each picking schedule chromosome in the picking schedule chromosome set, and the current optimal picking schedule chromosome is determined based on the objective function to generate the target picking schedule strategy. According to the filial generation inheritance probability of each picking schedule chromosome corresponding to the fitness value, two parental picking schedule chromosomes participating in filial generation inheritance are determined. The gene blocks in the two parental picking schedule chromosomes are recombined to generate two filial generation picking schedule chromosomes. The total backlog amount and the corresponding fitness value of the two filial generation picking schedule chromosomes are solved, and replacement determination is performed in combination with a preset annealing parameter, and a filial generation schedule chromosome set is generated based on the determination result. With a preset fixed number of iterations as the iteration termination condition, the recombination and replacement determination processes are iteratively executed. After the iteration is terminated, each picking schedule chromosome in the filial generation schedule chromosome set generated in the last iteration is decoded by using a preset mathematical programming model to determine the total backlog amount and the corresponding fitness value of each picking schedule chromosome in the picking schedule chromosome set, and the current optimal picking schedule chromosome is determined based on the objective function to generate the current picking schedule strategy.

[0052] In the present application, the current picking schedule strategy of the AGV is iteratively generated based on the picking efficiency of the pickers and the scheduling method. At the same time, the current picking schedule strategy of the AGV can be adjusted in real time according to the real-time picking efficiency transformation of the pickers. The current picking schedule strategy of the AGV can maintain adaptive change adjustment with the real-time picking efficiency of the pickers. Therefore, compared with the existing solutions, this method determines the optimal personnel scheduling method at the current moment with a fixed number of iterations for the real-time picking efficiency and scheduling method of the pickers, and then dynamically adjusts the current picking schedule strategy of the AGV based on this method, so as to avoid the problem that the picking efficiency of the pickers becomes the bottleneck restricting the overall efficiency of the system and improve the overall operation efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0054] Figure 1 FIG. 1 is a schematic diagram of an AGV distribution scenario provided for the background technology of the present application;

[0055] Figure 2 Flow chart of an AGV real-time transportation allocation method provided by an embodiment of the present application;

[0056] Figure 3 Schematic diagram of a picking scheduling chromosome provided by an embodiment of the present application;

[0057] Figure 4 Another schematic diagram of a picking scheduling chromosome provided by an embodiment of the present application;

[0058] Figure 5 Schematic diagram of a crossover operation provided by an embodiment of the present application;

[0059] Figure 6 Schematic diagram of a mutation operation provided by an embodiment of the present application;

[0060] Figure 7 Schematic diagram of the structure of an AGV real-time transportation allocation device exemplified by an embodiment of the present application. Detailed implementation manners

[0061] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0062] The present application can be used in many general-purpose or special-purpose computing device environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor devices, distributed computing environments including any of the above devices or equipment, and so on.

[0063] An embodiment of the present application provides an AGV real-time transportation allocation method. This method can be applied to various AGV real-time transportation systems or AGV warehousing systems, and can also be applied to various computer terminals or intelligent terminals. The execution subject can be the processor or server of the computer terminal or intelligent terminal.

[0064] Next, the solution of the present application will be introduced. The present application proposes the following technical solutions. For details, please refer to the following text.

[0065] Figure 2 Flow chart of an AGV real-time transportation allocation method disclosed by an embodiment of the present application. As Figure 2 shown, this method may include:

[0066] Step 1: Generate a set of picking scheduling chromosomes based on preset picking situation parameters.

[0067] Specifically, there are a preset number of picking schedule chromosomes generated by random permutation in the picking schedule chromosome set. The preset picking situation parameters include the total number of picking personnel in the warehousing center, product types, the total number of products to be sorted in each product type, the number of shifts, and the picking time for picking personnel to pick each product type, that is, the reciprocal of the skill level of picking personnel to pick each product type.

[0068] As Figure 3 shown, this application uses a 0-1 coding method to generate picking schedule chromosomes. "1" represents that the picking personnel go to work at shift t, and "0" represents that the picking personnel do not go to work at shift t. The work situation of picking personnel for each shift is coded, and a preset number of picking schedule chromosomes are randomly generated. Each picking schedule chromosome is a random permutation of 0 and 1. This application considers the skill level of picking personnel to pick different types of products in each work shift, and further considers the expected burden of continuous cross-shift work. Each picking schedule chromosome contains the scheduling situation of each personnel for all shifts.

[0069] Meanwhile, to further optimize the picking efficiency, this application specially designs the concept of "grouping". For example, if the number of groups is 2, it means that a group of picking workstations is arranged at each end of the warehouse. In this way, the AGV is first arranged to the group and then to the workstation. It should be noted that if the number of groups is 1, it is equivalent to not considering the grouping situation. At the same time, through different grouping arrangements of workers and workstations, the AGV path algorithm does not need to incur too much cost to achieve a strict match between products, workers, and workstations, but only needs to match to the corresponding group. As Figure 4 shown, the picking schedule chromosome can further carry group information. For example, a picking schedule chromosome can be composed of the scheduling situation of each personnel for each shift corresponding to all groups one by one. Each code in the picking schedule chromosome corresponds to the work situation of the picking personnel at shift t and the group situation to which the picking personnel belong.

[0070] Step 2: According to the picking efficiency of the current picking employees, use the preset mathematical programming model to decode each picking schedule chromosome in the picking schedule chromosome set, determine the total backlog volume of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the target picking schedule strategy.

[0071] Specifically, for a preset number of chromosomes in the randomly generated picking schedule chromosome set, first decode each picking schedule chromosome in the picking schedule chromosome set using a preset mathematical programming model, then determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set through the calculation formula of the total backlog, and determine the current optimal picking schedule chromosome through the objective function, that is, take the picking schedule chromosome with the smallest total backlog in the current picking schedule chromosome set as the current optimal picking schedule chromosome, and analyze the current optimal picking schedule chromosome to generate the corresponding target picking schedule strategy.

[0072] Based on the total backlog of each picking schedule chromosome in the picking schedule chromosome set, the fitness value corresponding to each picking schedule chromosome can also be calculated. The fitness value is the main indicator to describe the performance of an individual. According to the size of the fitness value, the individuals can be selected for survival of the fittest. The fitness value is the driving force for the genetic algorithm. The larger the individual fitness value, the better the individual, and the greater the chance of being inherited to the next generation. In this application, the larger the fitness value of the picking schedule chromosome, the higher the possibility of being selected to participate in the genetic inheritance of the offspring. On the contrary, the smaller the fitness value of the picking schedule chromosome, the lower the possibility of being selected to participate in the genetic inheritance of the offspring. The fitness value is used to determine two parental picking schedule chromosomes participating in the genetic inheritance of the offspring in the subsequent process.

[0073] Step 3: Calculate the genetic probability of the offspring of each picking schedule chromosome in the picking schedule chromosome set according to the fitness value of each picking schedule chromosome in the picking schedule chromosome set, and determine two parental picking schedule chromosomes participating in the genetic inheritance of the offspring in the picking schedule chromosome set according to the genetic probability of the offspring.

[0074] Specifically, the fitness function in the genetic algorithm is the main indicator to describe the performance of an individual. According to the size of the fitness, the individuals can be selected for survival of the fittest. The fitness is the driving force for the genetic algorithm. The larger the individual fitness value, the better the individual, and the greater the chance of being inherited to the next generation.

[0075] After calculating the fitness value of each picking schedule chromosome in the picking schedule chromosome set based on the fitness function, calculate the genetic probability of the offspring of each picking schedule chromosome, and select two parental picking schedule chromosomes participating in the genetic inheritance of the offspring in the picking schedule chromosome set according to the genetic probability of the offspring of each picking schedule chromosome. The operation of selecting the parental picking schedule chromosome is to select a certain number of individuals in the population and select those with better fitness to enter the offspring population. Common selection strategies include roulette wheel selection method, stochastic universal sampling method, tournament selection method, etc. It can be understood that the picking schedule chromosome with a larger fitness value has a greater possibility of being selected as the parental picking schedule chromosome, and the picking schedule chromosome with a smaller fitness value has a smaller possibility of being selected as the parental picking schedule chromosome.

[0076] Step 4: Recombine the gene blocks in the two parental selection and scheduling chromosomes according to the preset crossover probability and mutation probability to generate two offspring selection and scheduling chromosomes.

[0077] Specifically, after selecting the two parental selection and scheduling chromosomes participating in offspring inheritance, perform crossover determination and mutation determination respectively according to the preset crossover probability and mutation probability. If the result of the crossover determination is to perform a crossover transformation, then perform a crossover transformation recombination on the gene blocks in the two parental selection and scheduling chromosomes. If the result of the mutation determination is to perform a mutation determination, then perform a mutation transformation recombination on the gene blocks in the two parental selection and scheduling chromosomes. After recombination, two offspring selection and scheduling chromosomes are generated.

[0078] Step 5: Use the preset mathematical programming model to solve the total backlog and the corresponding fitness values of the two offspring selection and scheduling chromosomes.

[0079] Specifically, decode the two offspring selection and scheduling chromosomes using the preset mathematical programming model, and then determine the total backlogs corresponding to the two offspring selection and scheduling chromosomes respectively through the total backlog calculation formula. Further, determine the corresponding fitness values based on the total backlogs. The fitness values are used to replace and determine the generation of the offspring scheduling chromosome set.

[0080] Step 6: Perform replacement determination according to the fitness values of the two parental selection and scheduling chromosomes, the fitness values of the two offspring selection and scheduling chromosomes, and the preset annealing parameter, and generate the offspring scheduling chromosome set based on the determination result.

[0081] Specifically, since the fitness value can reflect the superiority of the chromosome, therefore, it is possible to determine the superiority levels of the two offspring selection and scheduling chromosomes after gene block recombination and the two parental selection and scheduling chromosomes before recombination according to the fitness values of the two parental selection and scheduling chromosomes and the fitness values of the two offspring selection and scheduling chromosomes, determine the more superior chromosome among them, and perform replacement determination based on the superiority levels and the preset annealing parameter, and finally obtain the offspring scheduling chromosome set.

[0082] Step 7: Use the preset fixed number of iterations as the iteration termination condition, and iteratively execute Steps 2 to 6. The offspring scheduling chromosome set generated in each iteration will participate in the iteration as the selection and scheduling chromosome set in the next iteration.

[0083] Specifically, in this application, the picking and scheduling chromosome set is genetically iterated with a preset fixed number of iterations as the iteration termination condition, and the picking and scheduling chromosome set is updated. In each iteration, the offspring scheduling chromosome set generated in the previous iteration will be used as the picking and scheduling chromosome set in this iteration, and the current optimal picking and scheduling chromosome will be determined based on the objective function to generate the target picking and scheduling strategy. Therefore, this application can determine the current optimal picking and scheduling chromosome corresponding to the current picking and scheduling chromosome set according to the picking efficiency of the picking personnel at the current moment, and determine the target picking and scheduling strategy corresponding to the current optimal picking and scheduling chromosome.

[0084] Step eight: After the iteration terminates, use the preset mathematical programming model to decode each picking and scheduling chromosome in the offspring scheduling chromosome set generated in the last iteration, determine the total backlog of each picking and scheduling chromosome in the picking and scheduling chromosome set, as well as the corresponding fitness value, and determine the current optimal picking and scheduling chromosome based on the objective function to generate the current picking and scheduling strategy.

[0085] Specifically, the objective function is:

[0086]

[0087] where RO j is the remaining number of products to be picked in picking group j, J is the total number of picking groups, and TRO is the total backlog of the picking and scheduling chromosome.

[0088] It can be understood that generally, as the number of iterations increases, the total backlog corresponding to the picking and scheduling chromosomes in the offspring scheduling chromosome set generated in each iteration will gradually decrease. After the last iteration is completed, the optimal picking and scheduling chromosome in the generated offspring scheduling chromosome set should be the optimal offspring scheduling chromosome generated in the entire iteration process, and the corresponding picking and scheduling strategy should be the optimal scheduling strategy among the picking and scheduling strategies generated in multiple iterations, that is, the picking and scheduling strategy with the smallest total backlog.

[0089] Step nine: Determine the current AGV transportation allocation strategy based on the current picking and scheduling strategy.

[0090] Specifically, after determining the current picking and scheduling strategy that can minimize the total backlog through the iterative algorithm, the current AGV transportation allocation strategy can be further determined based on the current picking and scheduling strategy. The current AGV transportation allocation strategy can change with the change of the current picking and scheduling strategy, that is, it can be adjusted according to the picking efficiency of the picking personnel.

[0091] When the picking efficiency of the picker becomes the bottleneck of the entire system, this application can improve the operating efficiency of the entire fulfillment system to a greater extent. Therefore, during the application process, a "self-diagnosis" step can be set to determine whether the picking efficiency of the picker is the operating bottleneck of the current warehouse center.

[0092] Generally speaking, it can be determined whether the bottleneck of the current system is the employee by detecting whether the current AGV is queuing because of the low efficiency of the picker, for example, by setting a queuing ratio threshold for the current AGV trolley (such as 50%). When the queuing ratio exceeds this threshold, we determine that the picking efficiency of the picker at this time is the operating bottleneck of the entire system, and at this time, the system operating efficiency can be greatly improved through this application.

[0093] It can be understood that after this application determines the randomly generated picking schedule chromosome set containing the picking staff scheduling information, it can iterate on the picking schedule chromosomes, determine the current picking schedule strategy corresponding to the current optimal picking schedule chromosome obtained after the iteration is completed, and correspondingly generate or adjust the current AGV transportation allocation strategy. Thus, considering the working efficiency of the picker, the allocation of product picking tasks is determined, and the transportation allocation strategy of the AGV is determined, achieving a substantial improvement in efficiency. In fact, for any picking schedule chromosome generated containing picking schedule information, regardless of its generation rule (including random generation, manual generation, or other methods, etc.), as long as the picking schedule chromosome is determined, this application can allocate picking tasks and AGVs based on corresponding parameters such as employee efficiency.

[0094] As can be seen from the above technical solutions, an AGV real-time transportation allocation method and device provided by an embodiment of the present application generate a picking schedule chromosome set through preset picking situation parameters. According to the picking efficiency of the current picking employees, each picking schedule chromosome in the picking schedule chromosome set is decoded by using a preset mathematical programming model to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, and the current optimal picking schedule chromosome is determined based on the objective function to generate the target picking schedule strategy. According to the filial generation inheritance probability of each picking schedule chromosome corresponding to the fitness value, two parental picking schedule chromosomes participating in filial generation inheritance are determined. The gene blocks in the two parental picking schedule chromosomes are recombined to generate two filial generation picking schedule chromosomes. The total backlog amount of the two filial generation picking schedule chromosomes and the corresponding fitness value are solved, and replacement determination is performed in combination with preset annealing parameters, and a filial generation schedule chromosome set is generated based on the determination result. With a preset fixed number of iterations as the iteration termination condition, the recombination and replacement determination processes are iteratively executed. After the iteration is terminated, each picking schedule chromosome in the filial generation schedule chromosome set generated in the last iteration is decoded by using a preset mathematical programming model to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, and the current optimal picking schedule chromosome is determined based on the objective function to generate the current picking schedule strategy.

[0095] In the present application, the current picking schedule strategy of the AGV is iteratively generated based on the picking efficiency of the pickers and the scheduling method. At the same time, the current picking schedule strategy of the AGV can be adjusted in real time according to the real-time picking efficiency change of the pickers. The current picking schedule strategy of the AGV can maintain adaptive change adjustment with the real-time picking efficiency of the pickers. Therefore, compared with the existing solutions, this method determines the optimal personnel scheduling method at the current moment with a fixed number of iterations for the real-time picking efficiency and scheduling method of the pickers, and then dynamically adjusts the current picking schedule strategy of the AGV based on this method, so as to avoid the problem that the picking efficiency of the pickers becomes the bottleneck restricting the overall system efficiency and improve the overall operation efficiency of the system.

[0096] In some embodiments of the present application, the process of the step two, in which each picking schedule chromosome in the picking schedule chromosome set is decoded by using a preset mathematical programming model according to the picking efficiency of the current picking employees to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, is introduced. Specifically, it may include:

[0097] S1. According to the picking efficiency of the picking employees, each picking schedule chromosome in the picking schedule chromosome set is decoded by using a preset mathematical programming model to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set.

[0098] Specifically, the calculation formula for determining the total backlog of each picking schedule chromosome in the picking schedule chromosome set is as follows:

[0099]

[0100]

[0101] where a ijn describes the duration allocated to product type n by each picker i in picking group j. N is the total number of product types, W is the total number of picking personnel, B n is the total number of products to be sorted in the nth product type, M in is the picking time for picker i to pick the nth product type, and is also the reciprocal of the skill level of picker i for picking the nth product type. RO j is the remaining number of products to be picked in picking group j. J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

[0102] S2. According to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set, determine the fitness value of each picking schedule chromosome in the picking schedule chromosome set.

[0103] Specifically, in this application, the objective function is to find the minimum value. In order to make the genetic factors evolve in the optimal direction during iteration, the fitness function in this application selects the reciprocal of the objective function, that is, the reciprocal of the total backlog of the picking schedule chromosome. The meaning expressed by the objective function is to minimize the total backlog.

[0104] The fitness function Fit(f(x)) is as shown in the following formula:

[0105]

[0106] where f(x) represents the total backlog of the picking schedule chromosome. For a feasible solution, the smaller the total backlog, the larger the fitness value, and the better the individual. Punish the generated infeasible individuals, that is, the fitness value of the infeasible solution is 0. Based on the total backlog of each picking schedule chromosome in the picking schedule chromosome set, the fitness of each individual can be obtained. It should be noted that when calculating the fitness value, it is necessary to first judge the existence of feasible solutions in the population, then calculate the individual fitness value, discard the individuals that do not meet the constraint conditions, and select the individuals with higher fitness values for inheritance.

[0107] The constraint conditions are:

[0108]

[0109] The constraint of the expression is that the total time allocated to each product is less than the total shift time.

[0110] Therefore, the process of determining the fitness value of each picking schedule chromosome in the picking schedule chromosome set is as follows:

[0111] ① Determine whether the total backlog of each picking schedule chromosome in the picking schedule chromosome set meets the constraint conditions;

[0112] ② Determine the fitness value of the picking schedule chromosome that meets the constraint conditions as the reciprocal of the total backlog of the picking schedule chromosome;

[0113] ③ Determine the fitness value of the picking schedule chromosome that does not meet the constraint conditions as zero.

[0114] In some embodiments of the present application, for step four, the process of recombining gene blocks in the two parental picking schedule chromosomes according to a preset crossover probability and mutation probability to generate two offspring picking schedule chromosomes is introduced. Specifically, it may include:

[0115] S1. Perform a crossover determination according to a preset crossover probability. If the result of the crossover determination is passed, randomly select two corresponding crossover points on the two parental picking schedule chromosomes for gene block crossover transformation. Otherwise, keep the two parental picking schedule chromosomes unchanged. The result generated after the crossover determination is two crossover-transformed schedule chromosomes.

[0116] Specifically, if the result of the crossover determination is passed, randomly select two corresponding crossover points on the two parental picking schedule chromosomes for gene block crossover transformation. If the result of the crossover determination is not passed, keep the two parental picking schedule chromosomes unchanged. The result generated after the crossover determination is two crossover-transformed schedule chromosomes. It can be understood that in the case of passing the determination, the two crossover-transformed schedule chromosomes are the chromosomes obtained after the crossover transformation. In the case of not passing the determination, the two crossover-transformed schedule chromosomes are the same as the two parental picking schedule chromosomes.

[0117] The crossover operation, also known as recombination, refers to the operation of replacing and recombining part of the structures of two parental individuals with a certain crossover probability to generate new individuals, which is one of the main operations for forming new individuals in the genetic algorithm. There are many methods for the crossover operation, such as one-point crossover, two-point crossover, multi-point crossover, uniform crossover, etc.

[0118] Taking the selection of two-point crossover for the crossover operation as an example, as Figure 5 shown, when the crossover determination is passed, randomly select two crossover points in the paired parental picking schedule chromosomes, and then perform the exchange of part of the gene blocks, that is, exchange the gene blocks in the middle area of the dotted line to obtain two crossover-transformed schedule chromosomes.

[0119] S2. Perform mutation determination according to a preset mutation probability. If the result of the mutation determination passes, randomly select a single gene block in the two cross-transformed scheduling chromosomes for replacement; otherwise, keep the two cross-transformed scheduling chromosomes unchanged. The result generated after the mutation determination is two mutated-transformed scheduling chromosomes.

[0120] Specifically, if the result of the mutation determination passes, randomly select a single gene block on the cross-transformed scheduling chromosome for replacement. If the result of the crossover determination fails, keep the two cross-transformed scheduling chromosomes unchanged. The result generated after the mutation determination is two mutated-transformed scheduling chromosomes. It can be understood that in the case of passing the determination, the two mutated-transformed scheduling chromosomes are the chromosomes obtained after the mutated transformation. In the case of failing the determination, the two mutated-transformed scheduling chromosomes are the same as the two cross-transformed scheduling chromosomes.

[0121] For the two cross-transformed scheduling chromosomes formed after the crossover operation, there is a certain probability that gene mutations will occur in the two cross-transformed scheduling chromosomes. The mutation operation also performs mutation determination based on the preset mutation probability. Similar to the biological world, the mutation probability is very low, and the typical value is between 0.001 and 0.1. Mutation provides opportunities for the generation of new individuals. Commonly used mutation operations include basic bit mutation, uniform mutation, boundary mutation, etc.

[0122] Taking the selection of basic bit mutation for the mutation operation as an example, as Figure 6 shown, in the case where the mutation determination passes, for one of the two cross-transformed scheduling chromosomes formed after the crossover operation, perform mutation transformation on the values of a certain bit or several bits of genes randomly specified with the mutation probability. In the legend, it is selected to swap a single gene, that is, replace a single gene on one of the two cross-transformed scheduling chromosomes.

[0123] S3. Use the two mutated-transformed scheduling chromosomes as the two offspring picking scheduling chromosomes.

[0124] In some embodiments of the present application, for the above step three, the process of calculating the offspring genetic probability of each picking scheduling chromosome according to the fitness value of each picking scheduling chromosome in the picking scheduling chromosome set and determining two parental picking scheduling chromosomes participating in offspring inheritance in the picking scheduling chromosome set according to the offspring genetic probability will be introduced. Specifically, it may include:

[0125] S1. Calculate the offspring genetic probability of each picking scheduling chromosome according to the fitness value of each picking scheduling chromosome in the picking scheduling chromosome set.

[0126] Specifically, the offspring inheritance probability is the proportion of the fitness value of each picking and scheduling chromosome in the total sum of the fitness values of all picking and scheduling chromosomes in the picking and scheduling chromosome set.

[0127] If the population size is L, that is, the total number of chromosomes in the picking and scheduling chromosome is L, and the fitness of an individual is Fit(x), that is, the fitness value of the picking and scheduling chromosome is Fit(x), then the offspring inheritance probability of the individual, that is, the picking and scheduling chromosome selected as the two parental picking and scheduling chromosomes participating in offspring inheritance is:

[0128] P(x i ) = Fit(x i ) / ∑Fit(x i ), (i = 1, 2,..., L).

[0129] S2. Adopt the roulette wheel strategy to determine two parental picking and scheduling chromosomes in the picking and scheduling chromosome set that participate in offspring inheritance based on the offspring inheritance probability of each picking and scheduling chromosome.

[0130] Specifically, roulette wheel selection calculates the probability of each individual appearing in the offspring based on the fitness value of the individual, and randomly selects individuals according to this probability to form the offspring population. The starting point of the roulette wheel selection strategy is that the greater the fitness value of an individual, the greater the probability of being selected.

[0131] When actually using the roulette wheel strategy, "cumulative probability" is often used for individual selection instead of selecting according to individual probability. The roulette wheel strategy adopted in this application is transformed into a maximization problem in the genetic algorithm, that is, solving the reciprocal of the original objective function. The cumulative probability q(x) of each individual, that is, each picking and scheduling chromosome, is:

[0132]

[0133] In some embodiments of this application, for step six, the process of determining replacement based on the fitness values of the two parental picking and scheduling chromosomes, the fitness values of the two offspring picking and scheduling chromosomes, and a preset annealing parameter, and generating an offspring scheduling chromosome set based on the determination result is introduced. Specifically, it may include:

[0134] S1. Determine the annealing probability according to the preset annealing parameter.

[0135] Specifically, the calculation formula for the annealing probability is:

[0136]

[0137] Among them, f(x) is the fitness value of the parent selection and scheduling chromosome, f′(x) is the fitness value of the offspring selection and scheduling chromosome, and T is the temperature.

[0138] S2. Compare the fitness values of the two parent selection and scheduling chromosomes and the fitness values of the two offspring selection and scheduling chromosomes, and determine the chromosome with the larger fitness value as the better chromosome.

[0139] S3. If the better chromosome is the offspring selection and scheduling chromosome, replace the parent selection and scheduling chromosome in the selection and scheduling chromosome set with the offspring selection and scheduling chromosome, and obtain the offspring scheduling chromosome set.

[0140] S4. If the better chromosome is the parent selection and scheduling chromosome, determine whether to replace it according to the annealing probability. If the determination result is yes, replace the parent selection and scheduling chromosome in the selection and scheduling chromosome set with the offspring selection and scheduling chromosome, and obtain the offspring scheduling chromosome set. Otherwise, keep the parent selection and scheduling chromosome in the selection and scheduling chromosome set unchanged, and use the selection and scheduling chromosome set as the offspring selection and scheduling chromosome.

[0141] Specifically, in order to effectively reduce gene loss and avoid falling into local optimality, the Metropolis criterion is embedded in this application. When the offspring selection and scheduling chromosome generated after the crossover operation and the mutation operation is superior to the parent selection and scheduling chromosome before the crossover operation and the mutation operation, fully accept the offspring selection and scheduling chromosome generated after the crossover operation and the mutation operation, and replace the parent selection and scheduling chromosome in the selection and scheduling chromosome set with the offspring selection and scheduling chromosome, so as to obtain the offspring scheduling chromosome set.

[0142] When the offspring selection and scheduling chromosome generated after the crossover operation and the mutation operation is inferior to the parent selection and scheduling chromosome before the crossover operation and the mutation operation, it is determined whether to replace it according to the annealing probability, that is, accept the offspring selection and scheduling chromosome generated after the crossover operation and the mutation operation with the annealing probability, and use it to replace the parent selection and scheduling chromosome in the selection and scheduling chromosome set. In addition, in this application, since the annealing temperature decays rapidly, the operation efficiency can be improved, and since the annealing temperature decays slowly, the solution can be more stable. Therefore, the annealing decay coefficient can be initially set to γ = 0.99 in this application.

[0143] Next, an AGV real-time transportation allocation device provided by an embodiment of this application will be described. The following description of an AGV real-time transportation allocation device can be mutually referred to with the above-described AGV real-time transportation allocation method.

[0144] See Figure 3 ,Figure 3 Schematic diagram of the structure of an AGV real-time transportation allocation device disclosed in an embodiment of the present application.

[0145] As Figure 3 shown, the device may include:

[0146] A chromosome generation unit 110, configured to generate a picking schedule chromosome set based on preset picking situation parameters, where there are a preset number of randomly arranged picking schedule chromosomes in the picking schedule chromosome set;

[0147] An iteration unit 120, configured to decode each picking schedule chromosome in the picking schedule chromosome set according to the picking efficiency of the current picking employee by using a preset mathematical programming model, determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set, and the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the target picking schedule strategy; calculate the offspring genetic probability of each picking schedule chromosome according to the fitness value of each picking schedule chromosome in the picking schedule chromosome set, and determine two parent picking schedule chromosomes participating in offspring inheritance in the picking schedule chromosome set according to the offspring genetic probability; recombine the gene blocks in the two parent picking schedule chromosomes according to the preset crossover probability and mutation probability to generate two offspring picking schedule chromosomes; use the preset mathematical programming model to solve the total backlog amount of the two offspring picking schedule chromosomes, and the corresponding fitness value; perform replacement determination according to the fitness values of the two parent picking schedule chromosomes, the fitness values of the two offspring picking schedule chromosomes, and the preset annealing parameter, and generate an offspring schedule chromosome set based on the determination result; use the preset fixed number of iterations as the iteration termination condition, and iteratively execute steps two to six. The offspring schedule chromosome set generated each time will participate in the iteration as the picking schedule chromosome set for the next iteration;

[0148] An allocation strategy unit 130, configured to, after the iteration terminates, decode each picking schedule chromosome in the offspring schedule chromosome set generated by the last iteration by using a preset mathematical programming model, determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set, and the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the current picking schedule strategy; determine the current AGV transportation allocation strategy based on the current picking schedule strategy.

[0149] As can be seen from the above technical solution, an AGV real-time transportation allocation device provided by an embodiment of the present application generates a picking schedule chromosome set through preset picking situation parameters. According to the picking efficiency of the current picking employees, a preset mathematical programming model is used to decode each picking schedule chromosome in the picking schedule chromosome set to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, and based on the objective function, the current optimal picking schedule chromosome is determined to generate the target picking schedule strategy. According to the offspring genetic probability of each picking schedule chromosome corresponding to the fitness value, two parent picking schedule chromosomes participating in offspring inheritance are determined. The gene blocks in the two parent picking schedule chromosomes are recombined to generate two offspring picking schedule chromosomes. The total backlog amount of the two offspring picking schedule chromosomes and the corresponding fitness value are solved, and a replacement determination is made in combination with the preset annealing parameter, and a set of offspring schedule chromosomes is generated based on the determination result. Using the preset fixed number of iterations as the iteration termination condition, the recombination and replacement determination processes are iteratively executed. After the iteration terminates, a preset mathematical programming model is used to decode each picking schedule chromosome in the set of offspring schedule chromosomes generated in the last iteration to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value, and based on the objective function, the current optimal picking schedule chromosome is determined to generate the current picking schedule strategy.

[0150] In the present application, the current picking schedule strategy of the AGV is iteratively generated based on the picking efficiency of the pickers and the scheduling method. At the same time, the current picking schedule strategy of the AGV can be adjusted in real time according to the real-time picking efficiency transformation of the pickers. The current picking schedule strategy of the AGV can maintain an adaptive change adjustment with the real-time picking efficiency of the pickers. Therefore, compared with the existing solutions, this method determines the optimal personnel scheduling method at the current moment with a fixed number of iterations for the real-time picking efficiency and scheduling method of the pickers, and then dynamically adjusts the current picking schedule strategy of the AGV based on this method, thereby avoiding the problem that the picking efficiency of the pickers becomes the bottleneck restricting the overall system efficiency and improving the overall operation efficiency of the system.

[0151] Optionally, the process of the iteration unit using a preset mathematical programming model to decode each picking schedule chromosome in the picking schedule chromosome set according to the picking efficiency of the current picking employees to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set and the corresponding fitness value may include:

[0152] According to the picking efficiency of the picking employees, a preset mathematical programming model is used to decode each picking schedule chromosome in the picking schedule chromosome set to determine the total backlog amount of each picking schedule chromosome in the picking schedule chromosome set;

[0153] Determine the fitness value of each picking schedule chromosome in the picking schedule chromosome set according to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set.

[0154] Optionally, the calculation formula for the iteration unit to determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set is:

[0155]

[0156]

[0157] where a ijn describes the duration assigned to product type n by each picker i in picking group j. N is the total number of product types, W is the total number of pickers, B n is the total number of products to be sorted in the nth product type, M in is the picking time for picker i to pick the nth product type, which is also the reciprocal of the skill level of picker i to pick the nth product type. RO j is the remaining number of products to be picked in picking group j. J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

[0158] Optionally, the objective function of the allocation strategy unit is:

[0159]

[0160] where RO j is the remaining number of products to be picked in picking group j. J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

[0161] Optionally, the process for the iteration unit to determine the fitness value of each picking schedule chromosome in the picking schedule chromosome set according to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set may include:

[0162] Determine whether the total backlog of each picking schedule chromosome in the picking schedule chromosome set meets the constraint conditions;

[0163] Determine the fitness value of the picking schedule chromosome that meets the constraint conditions as the reciprocal of the total backlog of the picking schedule chromosome;

[0164] Determine the fitness value of the picking schedule chromosome that does not meet the constraint conditions as zero.

[0165] Optionally, the process of the iteration unit recombining gene blocks in the two parental selection and scheduling chromosomes according to a preset crossover probability and mutation probability to generate two offspring selection and scheduling chromosomes may include:

[0166] Perform a crossover determination according to the preset crossover probability. If the result of the crossover determination is passed, randomly select two corresponding crossover points on the two parental selection and scheduling chromosomes for gene block crossover transformation; otherwise, keep the two parental selection and scheduling chromosomes unchanged. The result generated after the crossover determination is two crossover-transformed scheduling chromosomes.

[0167] Perform a mutation determination according to the preset mutation probability. If the result of the mutation determination is passed, randomly select a single gene block in the two crossover-transformed scheduling chromosomes for replacement; otherwise, keep the two crossover-transformed scheduling chromosomes unchanged. The result generated after the mutation determination is two mutation-transformed scheduling chromosomes.

[0168] Use the two mutation-transformed scheduling chromosomes as the two offspring selection and scheduling chromosomes.

[0169] Optionally, the process of the iteration unit calculating the offspring genetic probability of each selection and scheduling chromosome in the selection and scheduling chromosome set according to the fitness value of each selection and scheduling chromosome in the selection and scheduling chromosome set and determining two parental selection and scheduling chromosomes participating in offspring inheritance in the selection and scheduling chromosome set according to the offspring genetic probability may include:

[0170] Calculate the offspring genetic probability of each selection and scheduling chromosome according to the fitness value of each selection and scheduling chromosome in the selection and scheduling chromosome set. The offspring genetic probability is the proportion of the fitness value of each selection and scheduling chromosome in the sum of the fitness values of all selection and scheduling chromosomes in the selection and scheduling chromosome set.

[0171] Adopt a roulette wheel strategy to determine two parental selection and scheduling chromosomes participating in offspring inheritance in the selection and scheduling chromosome set based on the offspring genetic probability of each selection and scheduling chromosome.

[0172] Optionally, the process of the iteration unit performing a replacement determination according to the fitness values of the two parental selection and scheduling chromosomes, the fitness values of the two offspring selection and scheduling chromosomes, and a preset annealing parameter and generating an offspring scheduling chromosome set based on the determination result may include:

[0173] Determine an annealing probability according to the preset annealing parameter;

[0174] Compare the fitness values of the two parent picking and scheduling chromosomes and the fitness values of the two offspring picking and scheduling chromosomes, and determine the chromosome with the larger fitness value as the better chromosome;

[0175] If the better chromosome is the offspring picking and scheduling chromosome, then replace the parent picking and scheduling chromosome in the picking and scheduling chromosome set with the offspring picking and scheduling chromosome, and obtain an offspring scheduling chromosome set;

[0176] If the better chromosome is the parent picking and scheduling chromosome, then determine whether to replace it according to the annealing probability. If the determination result is yes, then replace the parent picking and scheduling chromosome in the picking and scheduling chromosome set with the offspring picking and scheduling chromosome, and obtain an offspring scheduling chromosome set. Otherwise, keep the parent picking and scheduling chromosome in the picking and scheduling chromosome set unchanged, and use the picking and scheduling chromosome set as the offspring picking and scheduling chromosome.

[0177] Optionally, the calculation formula for the annealing probability of the iteration unit is:

[0178]

[0179] where f(x) is the fitness value of the parent picking and scheduling chromosome, f′(x) is the fitness value of the offspring picking and scheduling chromosome, and T is the temperature.

[0180] Finally, it should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising a..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0181] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. The various embodiments can be combined with each other. Therefore, the present application will not be limited to these embodiments shown herein, but rather will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An AGV real-time transportation allocation method, characterized in that, it includes: Step 1: Generate a picking schedule chromosome set based on preset picking situation parameters. There are a preset number of picking schedule chromosomes randomly arranged in the picking schedule chromosome set; Step 2: According to the picking efficiency of the current picking employees, use a preset mathematical programming model to decode each picking schedule chromosome in the picking schedule chromosome set, determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set, and the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the target picking schedule strategy; Step 3: According to the fitness value of each picking schedule chromosome in the picking schedule chromosome set, calculate the offspring genetic probability of each picking schedule chromosome, and determine two parent picking schedule chromosomes participating in offspring inheritance in the picking schedule chromosome set according to the offspring genetic probability; Step 4: Recombine the gene blocks in the two parent picking schedule chromosomes according to the preset crossover probability and mutation probability to generate two offspring picking schedule chromosomes; Step 5: Use a preset mathematical programming model to solve the total backlog of the two offspring picking schedule chromosomes and the corresponding fitness value; Step 6: Make a replacement determination according to the fitness values of the two parent picking schedule chromosomes, the fitness values of the two offspring picking schedule chromosomes, and the preset annealing parameter, and generate an offspring schedule chromosome set based on the determination result; Step 7: Take the preset fixed number of iterations as the iteration termination condition, and iteratively execute Step 2 to Step 6. The offspring schedule chromosome set generated in each iteration will participate in the iteration as the picking schedule chromosome set in the next iteration; Step 8: After the iteration terminates, use a preset mathematical programming model to decode each picking schedule chromosome in the offspring schedule chromosome set generated in the last iteration, determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set, and the corresponding fitness value, and determine the current optimal picking schedule chromosome based on the objective function to generate the current picking schedule strategy; Step 9: Determine the current AGV transportation allocation strategy based on the current picking schedule strategy.

2. The method according to claim 1, characterized in that, in Step 2, according to the picking efficiency of the current picking employees, using a preset mathematical programming model to decode each picking schedule chromosome in the picking schedule chromosome set, determining the total backlog of each picking schedule chromosome in the picking schedule chromosome set, and the corresponding fitness value, includes: According to the picking efficiency of the picking employees, use a preset mathematical programming model to decode each picking schedule chromosome in the picking schedule chromosome set to determine the total backlog of each picking schedule chromosome in the picking schedule chromosome set; According to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set, determine the fitness value of each picking schedule chromosome in the picking schedule chromosome set.

3. The method according to claim 2, It is characterized in that The calculation formula for determining the total backlog of each picking schedule chromosome in the picking schedule chromosome set is: Among them, a ijn Describes the duration assigned to product type n by each picker i in picking group j. N is the total number of product types, W is the total number of pickers, and B n Is the total number of products to be sorted in the nth product type, and M in Is the picking time for picker i to pick the nth product type, and is also the reciprocal of the skill level of picker i to pick the nth product type. RO j Is the remaining number of products to be picked in picking group j. J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

4. The method according to claim 1, It is characterized in that In the step eight, the objective function is: Among them, RO j is the remaining quantity of products to be picked for picking group j, J is the total number of picking groups, and TRO is the total backlog of the picking schedule chromosome.

5. The method according to claim 2, It is characterized in that According to the fitness function and the total backlog of each picking schedule chromosome in the picking schedule chromosome set, determining the fitness value of each picking schedule chromosome in the picking schedule chromosome set includes: Determining whether the total backlog of each picking schedule chromosome in the picking schedule chromosome set meets the constraint conditions; Determining the fitness value of the picking schedule chromosome that meets the constraint conditions as the reciprocal of the total backlog of the picking schedule chromosome; Determining the fitness value of the picking schedule chromosome that does not meet the constraint conditions as zero.

6. The method according to claim 1, It is characterized in that In the step four, according to the preset crossover probability and mutation probability, recombining the gene blocks in the two parental picking schedule chromosomes to generate two offspring picking schedule chromosomes, including: Performing a crossover determination according to the preset crossover probability. If the result of the crossover determination is passed, randomly select two corresponding crossover points on the two parental picking schedule chromosomes for gene block crossover transformation, otherwise keep the two parental picking schedule chromosomes unchanged. The result generated after the crossover determination is two crossover transformation schedule chromosomes; Performing a mutation determination according to the preset mutation probability. If the result of the mutation determination is passed, randomly select a single gene block in the two crossover transformation schedule chromosomes for replacement, otherwise keep the two crossover transformation schedule chromosomes unchanged. The result generated after the mutation determination is two mutation transformation schedule chromosomes; Taking the two mutation transformation schedule chromosomes as the two offspring picking schedule chromosomes.

7. The method according to claim 1, It is characterized in that In the step three, according to the fitness value of each picking schedule chromosome in the picking schedule chromosome set, calculating the offspring inheritance probability of each picking schedule chromosome, and determining two parental picking schedule chromosomes participating in offspring inheritance in the picking schedule chromosome set according to the offspring inheritance probability, including: According to the fitness value of each picking schedule chromosome in the picking schedule chromosome set, calculating the offspring inheritance probability of each picking schedule chromosome. The offspring inheritance probability is the proportion of the fitness value of each picking schedule chromosome in the sum of the fitness values of all picking schedule chromosomes in the picking schedule chromosome set; Adopting the roulette wheel strategy, based on the offspring inheritance probability of each picking schedule chromosome, determining two parental picking schedule chromosomes participating in offspring inheritance in the picking schedule chromosome set.

8. The method according to claim 1, It is characterized in that In the step six, according to the fitness values of the two parental picking schedule chromosomes, the fitness values of the two offspring picking schedule chromosomes, and the preset annealing parameter, performing a replacement determination, and generating an offspring schedule chromosome set based on the determination result, including: Determine the annealing probability according to the preset annealing parameters; Compare the fitness values of the two parent picking and scheduling chromosomes and the fitness values of the two offspring picking and scheduling chromosomes, and determine the chromosome with the larger fitness value as the better chromosome; If the better chromosome is the offspring picking and scheduling chromosome, replace the parent picking and scheduling chromosome in the picking and scheduling chromosome set with the offspring picking and scheduling chromosome, and obtain the offspring scheduling chromosome set; If the better chromosome is the parent picking and scheduling chromosome, determine whether to replace according to the annealing probability. If the determination result is yes, replace the parent picking and scheduling chromosome in the picking and scheduling chromosome set with the offspring picking and scheduling chromosome, and obtain the offspring scheduling chromosome set. Otherwise, keep the parent picking and scheduling chromosome in the picking and scheduling chromosome set unchanged, and use the picking and scheduling chromosome set as the offspring picking and scheduling chromosome.

9. The method according to claim 8, wherein, the calculation formula of the annealing probability is: where f(x) is the fitness value of the parent picking and scheduling chromosome, f′(x) is the fitness value of the offspring picking and scheduling chromosome, and T is the temperature.

10. An AGV real-time transportation allocation device, wherein, comprising: a chromosome generation unit, configured to generate a picking and scheduling chromosome set based on preset picking situation parameters, where there are a preset number of picking and scheduling chromosomes randomly arranged in the picking and scheduling chromosome set; an iteration unit, configured to, according to the picking efficiency of the current picking employee, use a preset mathematical programming model to decode each picking and scheduling chromosome in the picking and scheduling chromosome set, determine the total backlog amount of each picking and scheduling chromosome in the picking and scheduling chromosome set and the corresponding fitness value, and determine the current optimal picking and scheduling chromosome and generate the target picking and scheduling strategy based on the objective function; calculate the offspring genetic probability of each picking and scheduling chromosome according to the fitness value of each picking and scheduling chromosome in the picking and scheduling chromosome set, and determine two parent picking and scheduling chromosomes participating in offspring inheritance in the picking and scheduling chromosome set according to the offspring genetic probability; recombine the gene blocks in the two parent picking and scheduling chromosomes according to the preset crossover probability and mutation probability to generate two offspring picking and scheduling chromosomes; use the preset mathematical programming model to solve the total backlog amount of the two offspring picking and scheduling chromosomes and the corresponding fitness value; perform replacement determination according to the fitness values of the two parent picking and scheduling chromosomes, the fitness values of the two offspring picking and scheduling chromosomes, and the preset annealing parameters, and generate an offspring scheduling chromosome set based on the determination result; use the preset fixed number of iterations as the iteration termination condition, and iteratively execute steps two to six. The offspring scheduling chromosome set generated each time will participate in the iteration as the picking and scheduling chromosome set for the next iteration; A distribution strategy unit is used to, after the iteration terminates, decode each picking scheduling chromosome in the offspring scheduling chromosome set generated in the last iteration by using a preset mathematical programming model, determine the total backlog of each picking scheduling chromosome in the picking scheduling chromosome set, as well as the corresponding fitness value, and determine the current optimal picking scheduling chromosome based on the objective function to generate the current picking scheduling strategy; Based on the current picking scheduling strategy, determine the current AGV transportation distribution strategy.

Images