A Two-stage Scheduling Method for Logistics in Intelligent Manufacturing Workshop Based on Genetic Algorithm

Abstract

An embodiment of the present invention discloses a two-stage scheduling method for the logistics in an intelligent manufacturing workshop based on a genetic algorithm, which relates to the technical field of intelligent manufacturing. The present invention includes: establishing a two-stage optimization model of task batching and path optimization with the shortest total picking distance as the objective function, calling a genetic algorithm for the research problem, nesting a path strategy for a corridor-type warehouse for solution, designing crossover and mutation strategies suitable for this problem to improve the solution accuracy, and being able to automatically realize order batching, so that the total picking distance is the shortest. Thus, the two-stage scheduling problem of "task batching - path optimization" in the logistics of an intelligent manufacturing workshop is solved, thereby optimizing the order batching scheduling method in the intelligent workshop logistics and improving the comprehensive production efficiency of the intelligent workshop.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent manufacturing, and particularly relates to a two-stage scheduling method for intelligent manufacturing workshop logistics based on a genetic algorithm. Background Art

[0002] With the development of intelligent manufacturing technology, the production efficiency and production quality of production lines have been significantly improved. However, in the actual production process, the material delivery efficiency of the intelligent logistics transportation system that supplies raw materials to the production line will restrict the comprehensive production efficiency of the manufacturing industry. If the raw material distribution is not timely, the production workshop may suspend or interrupt production, which is a major production accident in actual production. Therefore, this poses a huge challenge to the intelligent manufacturing workshop logistics transportation system, and how to improve the production material delivery efficiency of the production system has become a pain point problem to be solved urgently.

[0003] The material delivery process of the intelligent logistics transportation system can be specifically described as follows: during a certain period of time, the production system has multiple dynamically arriving orders that need to be produced, and the production system periodically issues tasks to the intelligent logistics transportation system in the form of orders. Each order includes information such as the item and quantity of production raw materials; according to the order information, the raw material warehouse completes the picking task; after the picking is completed and sorted and packed, it is transported to the production line. Specifically, the core of improving the material delivery efficiency of the intelligent logistics transportation system lies in the picking link, which includes two major modules: order batching scheduling and picking path optimization. Among them, order batching scheduling is the premise of path optimization and the key to picking operation optimization. Due to the particularity of warehousing logistics, some existing picking path strategies, such as the S-shaped path strategy and the midpoint strategy, have been widely applied. Under the S-shaped path picking path strategy, if the goods to be picked are in the first row of the shelf in a certain column of shelves, walking according to the S-shaped path will increase unnecessary walking distance. Obviously, order batching and picking path optimization are interrelated. Each batching result that meets the constraint conditions corresponds to a total picking distance, and as the batching result changes, the total picking distance changes accordingly. Therefore, the order batching strategy will have a great impact on the total picking distance. Therefore, the core problem of improving the efficiency of the production material delivery system lies in how to reasonably optimize the two-stage scheduling problem of "task batching - path optimization".

[0004] However, so far, due to the high investment cost of automated systems, many traditional manufacturing enterprises in our country still use experience-based manual order batching scheduling. Obviously, this method can no longer meet the requirements of high-speed production lines for delivery efficiency. At present, many optimization models of order batching optimization methods focus on minimizing the number of bin outings. This method cannot be well combined with the path optimization strategy and cannot obtain the overall optimal result of order batching and picking path optimization. In addition, due to the particularity of the production system, there are many types and large quantities of materials, and the existing order batching methods also need to be further optimized and improved.

[0005] Therefore, how to further optimize the current order batching scheduling method in the intelligent workshop logistics to improve the comprehensive production efficiency of the intelligent workshop has become a research topic. Summary of the Invention

[0006] An embodiment of the present invention provides a two-stage scheduling method for intelligent manufacturing workshop logistics based on genetic algorithm, which can optimize the order batching scheduling method in intelligent workshop logistics and improve the comprehensive production efficiency of intelligent workshops.

[0007] To achieve the above object, the embodiment of the present invention adopts the following technical solutions:

[0008] In the first aspect, the method provided by the embodiment of the present invention includes:

[0009] S1. Obtain the order demand information of the production system of the intelligent manufacturing workshop;

[0010] S2. According to the preset path planning strategy and the order demand information, establish an objective function for the corresponding order picking task;

[0011] S3. Establish an order batching and picking path planning model associated with the objective function;

[0012] S4. Use the genetic algorithm to obtain the optimization result of order batching and picking path through the order batching and picking path planning model.

[0013] In the second aspect, the two-stage scheduling device for intelligent manufacturing workshop logistics based on genetic algorithm of the present invention includes:

[0014] An order collection module, configured to obtain the order demand information of the production system of the intelligent manufacturing workshop;

[0015] A preprocessing module, configured to establish an objective function for the corresponding order picking task according to the preset path planning strategy and the order demand information;

[0016] An order batching and picking path planning module, configured to establish an order batching and picking path planning model associated with the objective function;

[0017] A processing module, configured to use a genetic algorithm to obtain an optimized result of order batching and picking path through the order batching and picking path planning model.

[0018] The two-stage scheduling method for intelligent manufacturing workshop logistics based on genetic algorithm provided by the embodiment of the present invention establishes a two-stage optimization model of task batching and path optimization with the shortest total picking distance as the objective function, calls the genetic algorithm for the research problem, nests the path strategy of the aisle-type warehouse for solution, designs the crossover and mutation strategies suitable for this problem, improves the solution accuracy, and can automatically realize order batching, making the total picking distance the shortest. Thus, it solves the two-stage scheduling problem of "task batching - path optimization" in intelligent manufacturing workshop logistics, optimizes the order batching scheduling method in intelligent workshop logistics, and improves the comprehensive production efficiency of intelligent workshops. Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0020] Figure 1 It is a schematic flowchart of a specific example provided by the embodiment of the present invention;

[0021] Figure 2 It is an algorithm flowchart of a specific example provided by the embodiment of the present invention;

[0022] Figure 3 It is a warehouse layout diagram of a specific example provided by the embodiment of the present invention;

[0023] Figure 4 It is an optimal batching scheme and path planning diagram of a specific example provided by the embodiment of the present invention;

[0024] Figure 5 It is a genetic algorithm iteration diagram of a specific example provided by the embodiment of the present invention;

[0025] Figure 6 It is a schematic flowchart of the method provided by the embodiment of the present invention;

[0026] Figure 7 It is a schematic diagram of the device structure provided by the embodiment of the present invention. Detailed Embodiments

[0027] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention will be described in detail hereinafter. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary only for explaining the present invention and should not be construed as limiting the present invention. Those skilled in the art of the present technology can understand that unless specifically stated otherwise, the singular forms "a", "an", "the" and "said" used herein may also include the plural forms. It should be further understood that the term "comprising" used in the specification of the present invention means the presence of the described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The phrase "and / or" used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art of the present technology can understand that unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as the general understanding of those of ordinary skill in the field to which the present invention belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless defined as herein.

[0028] In this embodiment, an optimization model is established for the two-stage scheduling problem of "task batching - path optimization", a mathematical model with the minimum total picking distance as the objective, and the gurobi solver is used to verify the effectiveness of the model. However, with the increase in the data scale of the actual problem, the solving efficiency of the gurobi solver obviously fails to meet the requirements of actual production. Therefore, the present invention designs a genetic algorithm for solving according to the actual problem to achieve the overall optimum of a large number of order batches.

[0029] An embodiment of the present invention provides a two-stage scheduling method for the logistics of an intelligent manufacturing workshop based on a genetic algorithm, as Figure 6 shown, including:

[0030] S1. Obtain the order demand information of the production system of the intelligent manufacturing workshop;

[0031] S2. According to the preset path planning strategy and the order demand information, establish an objective function for the corresponding order picking task;

[0032] S3. Establish an order batching and picking path planning model associated with the target function;

[0033] S4. Use the genetic algorithm to obtain the optimization results of order batching and picking paths through the order batching and picking path planning model.

[0034] In this embodiment, the traditional genetic algorithm is improved for the research problem background; based on the storage location information and layout diagram of the corridor-style raw material warehouse, with the shortest total picking distance as the target function and the order information as the guide. By designing a hybrid crossover strategy, on the basis of elite retention, a local search is performed on some elite chromosomes of each generation with a certain probability to improve the convergence speed and solution accuracy of the algorithm. Specifically, in each order batching and path planning, the order sequence to be processed is shuffled according to certain rules each time to generate a legal sequence, an initial population is generated, and these order sequences are batched according to certain strategies to determine the nodes to be passed through for each batch, and the S-shaped path strategy algorithm is called for all batches to calculate their fitness values; then the population is continuously selected, and at the same time, the newly designed crossover and mutation operations are completed on the selected population. Finally, the optimal order batching plan is planned through continuous iteration.

[0035] In this embodiment, in S1, it includes: real-time collecting the order demand information of the production system within a specified time period through information technology. Among the order demand information, at least include: raw material item number information and demand quantity information. Among them, information technology refers to information exchange through computers, enterprise resource planning (ERP) systems, etc.

[0036] In this embodiment, the distance between two shelves is solved according to the warehouse layout, and the target function corresponding to the order picking task is established:

[0037]

[0038] Among them, d ij represents the distance between node i and node j, x ijdb represents a variable with a value between 0 and 1. When passing through arc (i, j) in batch b of sorting station d, x ijdb = 1, otherwise x ijdb = 0, z represents the total distance of all the goods required for sorting all orders, A represents the set of arcs between any two nodes, d ∈ V D , b ∈ B d , V D represents the set of all sorting stations, B d represents the set of batches for each sorting station; the preset path planning strategy adopts the S-shaped path strategy.

[0039] For example Figure 3The warehouse layout shown includes two sorting areas and 50 shelves. The sorting areas of the warehouse are the starting and ending points of the AGVs. The warehouse contains a total of 10 columns of shelves. There are single-row shelves on both sides of the warehouse, and the rest are back-to-back shelves. Each shelf has the same height. The width of the longitudinal aisle in the warehouse is 1m, and the width of the transverse aisle is 2m. The two types of aisles are numbered respectively. Since the AGV travels along the aisle and the aisles are arranged in the horizontal and vertical directions, the travel distance of the AGV is the broken-line distance. In this problem, the S-shaped path strategy is adopted for the travel strategy. Define the shelves and stations in the form of classes, and perform object instantiation operations on each shelf. The information stored includes: the number of the shelf; the aisle information where the shelf is located; the coordinates of the shelf; the type of the shelf; the types of items on the shelf; the quantity of items on the shelf. In the warehouse, use d ij to represent the distance between node i and node j.

[0040] In this embodiment, in S3, it includes: establishing an order batching and picking path planning model with the shortest total picking distance as the objective function; among them, the positions of each sorting station, shelf and the goods information are known, the number and capacity of AGVs are known. When multiple production order demands are generated within one hour, it is required to reasonably plan the order batching method and the picking path to minimize the objective function and meet the constraint conditions. The constraint conditions of the objective function include: one AGV is responsible for the picking task of one batch, starting from the sorting station and returning to the same sorting station; and, all the goods in each batch responsible for by each AGV are less than or equal to its maximum capacity; load balancing of the sorting stations.

[0041] Furthermore, before running the order batching and picking path planning model, it also includes: determining the batch set B of each sorting station d , where:

[0042]

[0043] o represents the goods in the order, O represents the set of all goods to be picked, k represents the goods taken out, M represents the set of all goods that need to be picked, w ok represents the goods in order o, V D represents the set of all sorting stations, and Q represents the maximum load of one batch / AGV.

[0044] In the order batching and picking path planning model of this embodiment, it includes:

[0045] Allocate each kind of goods in each order to one batch;

[0046] Constrain the capacity of the batches to which the goods are allocated, including: ensuring that all the goods in each order can only be picked in one batch and the order cannot be split, and, each kind of goods to be picked in each batch is only picked on one shelf;

[0047] For each batch, if item k is picked from shelf s, then shelf s must be visited.

[0048] Generate a corresponding loop for each batch. In the generated loop, the AGV responsible for the batch returns to the sorting station after leaving the sorting station, and the remaining load of the AGV is greater than or equal to the total weight of the items to be picked on the shelf before reaching the shelf to be visited.

[0049] Specifically: Through sub-models (2) and (3), each type of item in each order must and can only be assigned to one batch.

[0050] Constrain the capacity of each batch through sub-model (4).

[0051] Constrain that all items in each order can only be picked in one batch through sub-model (5), where the order is indivisible.

[0052] Determine all items to be picked in each batch through sub-model (6).

[0053] Constrain that each type of item to be picked in each batch can only be picked on one shelf through sub-model (7).

[0054] Process each batch through sub-models (8) and (9). Among them: If item k is picked from shelf s, then shelf s must be visited.

[0055] Represent the composition of the line graph and form a loop through sub-model (10).

[0056] Constrain that each batch can only form one loop through sub-model (11).

[0057] Constrain that the load starting from the sorting station is 0 through sub-model (12).

[0058] Constrain that the remaining load is greater than or equal to the total weight of the items to be picked on the shelf before reaching the shelf to be visited through sub-model (13).

[0059] Constrain that the load between shelves that each batch does not need to visit is 0 through sub-model (14).

[0060] Eliminate sub-loops through sub-model (15).

[0061] Among them, sub-models (2) to (15) include:

[0062]

[0063] Among them, V represents the set of all points, V SLet \(V\) be the set of all shelves (at least including one item to be picked). D Let \(S\) denote the set of all sorting stations; \(A\) represents the set of arcs between any two nodes \(i\) and \(j\); \(M\) represents the set of all items to be picked. Let \(V_k\) denote the set of shelves containing item \(k\in M\); \(O\) represents the set of all orders. o Let \(M_o\) denote the set of all items to be picked for order \(o\in O\). d Let \(B\) denote the set of batches for each sorting station; \(Q\) represents the maximum load of a batch / AGV. ok Let \(w_{k,o}\) denote the weight of item \(k\) in order \(o\), where \(k\in M\). ksdb Let \(z_{k,b}\) be a 0 - 1 variable. In batch \(b\in B\), if item \(k\) is picked from shelf \(v\in V\), then \(z_{k,b}=1\); otherwise, \(z_{k,b}=0\). d In batch \(b\in B\), if item \(k\) is picked from shelf \(v\in V\), then \(z_{k,b}=1\); otherwise, \(z_{k,b}=0\). If item \(k\) is picked from shelf \(v\in V\), then \(z_{k,b}=1\); otherwise, \(z_{k,b}=0\). psrb If item \(k\) is picked from shelf \(v\in V\), then \(z_{k,b}=1\); otherwise, \(z_{k,b}=0\). psrb If item \(k\) is picked from shelf \(v\in V\), then \(z_{k,b}=1\); otherwise, \(z_{k,b}=0\). kodb Let \(u_{k,d,b}\) be a 0 - 1 variable. If item \(k\) in order \(o\in O\) is picked in batch \(b\in B\) at sorting station \(d\in S\), then \(u_{k,d,b}=1\); otherwise, \(u_{k,d,b}=0\). d If item \(k\) in order \(o\in O\) is picked in batch \(b\in B\) at sorting station \(d\in S\), then \(u_{k,d,b}=1\); otherwise, \(u_{k,d,b}=0\). kodb If item \(k\) in order \(o\in O\) is picked in batch \(b\in B\) at sorting station \(d\in S\), then \(u_{k,d,b}=1\); otherwise, \(u_{k,d,b}=0\). kodb If item \(k\) in order \(o\in O\) is picked in batch \(b\in B\) at sorting station \(d\in S\), then \(u_{k,d,b}=1\); otherwise, \(u_{k,d,b}=0\). ijdb Let \(y_{i,j}\) be a continuous variable representing the load on arc \((i,j)\). Further:

[0064]

[0065] In this embodiment, based on the genetic algorithm, Python is used to perform a simulation analysis on the total picking distance objective function model to obtain the shortest total picking distance. For example, Figure 2 As shown, the solution process based on the genetic algorithm is as follows:

[0066] (1) Determine the optimization objective and establish the optimization model.

[0067] (2) Determine the coding method.

[0068] Each chromosome \(C\) adopts a real - number coding method. The expression of \(C\) is \(C = [O_1,O_2,\cdots,O_n]\), where the number of genes corresponds to the order number. For example, \(C = [2,5,6,8,9,1,4,3,7,10]\) represents 10 orders respectively.

[0069] (3) Determine the population size.

[0070] (4) Determine the termination condition and the fitness function.

[0071] The termination condition is to judge whether it is greater than the maximum number of iterations Gen. Since the initial coding does not batch the orders, batch operation needs to be performed in the decoding process. Aiming to achieve the load balance of each sorting station and meet the AGV capacity limit condition at the same time. Traverse each order to complete the order batching operation. After obtaining the batching result, for each batch, a picking list is formed, that is, the set of goods to be picked for each batch. The shelves to be picked for this batch are determined by the set of goods, and then the picking path distance of this batch is calculated according to the S-shaped path strategy. Finally, the fitness value is obtained, and its calculation formula is:

[0072]

[0073] (5) Initialize the population and conduct fitness evaluation

[0074] (6) Judge whether the termination condition is met. If it is met, generate the optimal batching result and the total picking distance S. Otherwise, execute (5).

[0075] (7) Genetic operation. According to the previous genetic control parameters, randomly select individuals in the sub-population for crossover and mutation operations, and perform the optimal preservation operation on the best individual of the previous generation to generate the next generation. The crossover operation is specifically as follows: On the premise that a population that meets the constraint conditions has been constructed, crossover is performed on the two parent chromosomes parent1 and parent2 selected according to the selection operation. Randomly find the sub-order sequences A and B with the same index values [a, b] on the two chromosomes, sort the other sequences in parent1 that do not contain the sub-order sequence B to generate sequence A_1; at the same time, perform the same operation on parent2 to generate sequence B_1, and then perform crossover according to the index position. For example: The chromosome encodings of the two parent individuals are A = [2, 5, 8, 6, 9, 1, 4, 3, 7, 10] and B = [6, 4, 7, 8, 5, 2, 1, 10, 3, 9] respectively. Randomly select two index values [0, 3], then the offspring individuals are A' = [6, 4, 7, 2, 5, 8, 9, 3, 7, 10] and B' = [2, 5, 8, 6, 4, 7, 1, 10, 3, 9]. Set the mutation probability to prob = 0.1. The mutation operation is specifically as follows: Traverse the population of parent chromosomes, generate a random number between 0 and 1 each time. If the random number is less than the mutation probability, perform the mutation operation; when mutating, randomly select two index values [a, b] of the chromosome, and then exchange the values at their corresponding index positions. For example: The encoding of a certain chromosome is [2, 5, 6, 8, 9, 1, 4, 3, 7, 10]. Randomly select the positions of its indexes 2 and 8 for exchange, then the mutated chromosome is [2, 5, 7, 8, 9, 1, 4, 3, 6, 10].

[0076] Optimization Simulation Example Analysis of the Solution in the "Task Batching - Route Optimization" Two-Stage Scheduling Scenario:

[0077] Suppose the layout of a single-zone aisle-type warehouse in a certain distribution center is as Figure 3 shown, with two identical horizontal aisles and five identical vertical aisles. The horizontal aisles are front-aisle and back-aisle respectively, and the vertical aisles are Aisle 1 - 5 respectively. Each shelf is numbered separately. The specific parameters of this warehouse are set as shown in Table 1.

[0078] Table 1. Warehouse Parameters

[0079] Parameter Parameter description Parameter value d Longitudinal aisle width 1m k Front-aisle width 2m w Shelf width 1m l Shelf length 2m

[0080] (1) Batch picking in the order of order arrival

[0081] Taking 10 orders as an example, the load of each batch is 2 unit. Sorting according to the order of order arrival, each batch does not exceed the batch load, and it is divided into 4 batches. The picking route adopts the S-shaped route strategy, and the total picking distance is 144m. The picking data is shown in Table 2:

[0082] Table 2. Results of Picking in the Order of Orders

[0083]

[0084] (2) Optimized picking with "Task Batching - Route Optimization" two-stage scheduling

[0085] Taking the same orders as an example, perform "Task Batching - Route Optimization" two-stage scheduling on them, and call the genetic algorithm to solve. The genetic algorithm (Genetic Algorithm, GA) of the present invention sets the crossover probability and mutation probability to p c = 0.9, p m = 0.1, the population size Np = 20, and the number of iterations Gen = 200.

[0086] The calculated optimal order batching results are shown in Table 3, the total picking route is 88m, and the picking route of each batch is as Figure 4 shown.

[0087] Table 3. Results of Optimized Picking with "Task Batching - Route Optimization" Two-Stage Scheduling

[0088]

[0089] To analyze the convergence of the genetic algorithm, based on 10 sets of operation data, calculate the maximum and minimum values of each iteration of each operation to make an iteration graph. The iteration graph of the genetic algorithm is as Figure 5As shown in the figure. The results show that as the number of iterations increases, the total picking distance continuously decreases. After the number of iterations reaches 100 times, the maximum and minimum values of the total picking distance stabilize at a fixed value. Therefore, this algorithm has better convergence. At the same time, compared with the picking method of batching according to the order arrival sequence, the total picking distance is reduced by 38.89%, greatly improving the picking efficiency.

[0090] In practical applications, this embodiment can combine the actual situation of the enterprise's intelligent logistics transportation system to construct an order batching optimization model suitable for the intelligent logistics picking system, which has good guiding significance for the internal logistics distribution and planning of the enterprise. By constructing appropriate crossover operators and mutation operators, the advantages of good adaptability and convergence of the genetic algorithm are well applied to this problem, solving the order batching optimization problem in the enterprise's intelligent logistics transportation system, ensuring the optimization effect, and at the same time creating conditions for reducing logistics costs, improving the overall production efficiency of the enterprise, and improving the production benefits of the enterprise.

[0091] This embodiment also designs a two-stage scheduling device for the logistics of an intelligent manufacturing workshop based on the genetic algorithm. This device can run on the server of the workshop logistics control center in the intelligent manufacturing workshop, as Figure 7 shown. This device includes:

[0092] An order acquisition module, used to obtain the order demand information of the production system of the intelligent manufacturing workshop;

[0093] A preprocessing module, used to establish an objective function for the corresponding order picking task according to the preset path planning strategy and the order demand information;

[0094] An order batching and picking path planning module, used to establish an order batching and picking path planning model associated with the objective function;

[0095] A processing module, used to obtain the optimized results of order batching and picking paths through the order batching and picking path planning model by using the genetic algorithm.

[0096] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A two-stage scheduling method for the logistics of an intelligent manufacturing workshop based on a genetic algorithm, characterized in that, Including: S1. Obtain the order demand information of the production system in the intelligent manufacturing workshop; S2. According to the preset path planning strategy and the order demand information, establish an objective function for the corresponding order picking task; S3. Establish an order batching and picking path planning model associated with the objective function; S4. Use the genetic algorithm to obtain the optimization results of order batching and picking paths through the order batching and picking path planning model; In S3, it includes: Taking the shortest total picking distance as the objective function, establish an order batching and picking path planning model; The constraint conditions of the objective function include: one AGV is responsible for the picking task of one batch, starting from the sorting station and returning to the same sorting station; and, all the goods in each batch responsible for by one AGV are less than or equal to its maximum capacity; In the order batching and picking path planning model, it includes: Allocate each type of goods in each order to one batch; Constrain the capacity of the batches allocated with goods, including: ensuring that all the goods in each order can only be picked in one batch and the order cannot be split, and, each type of goods to be picked in each batch is only picked on one shelf; In the order batching and picking path planning model, it includes: For each batch, if goods k are picked from shelf s, then this shelf s must be visited; Generate a corresponding loop for each batch. Among them, in the generated loop, the AGV responsible for this batch will return to the sorting station after leaving the sorting station, and, the remaining load of the AGV before reaching the shelf to be visited is greater than or equal to the total weight of the items to be picked on this shelf; In the order batching and picking path planning model, it includes: sub-models (2) to (7); Through sub-models (2) and (3), all types of goods in each order must and can only be allocated to one batch; Constrain the capacity of the batches through sub-model (4); Constrain that all the goods in each order can only be picked in one batch through sub-model (5), where the order cannot be split; Determine all the goods to be picked in each batch through sub-model (6); Constrain that each type of goods to be picked in each batch can only be picked on one shelf through sub-model (7); Among them, sub-models (2) to (7) include: Among them, V represents the set of all points, V S is the set of shelves of all shelves, V D represents the set of all sorting stations; A represents the set of arcs between any two nodes i and node j, M is the set of all goods to be picked, represents the set of shelves containing item k ∈ M; O represents the set of all orders, M o represents the set of all goods to be picked in an order o in the order set O, o ∈ O, B d represents the set of batches at each sorting station; Let \(Q\) denote the maximum load of a batch / AGV, \(w\) ok denotes the weight of item \(k\) in order \(o\), \(k\in M\); \(z\) ksdb is a 0 - 1 variable. In batch \(b\in B\) d item \(k\) is picked up from the shelf ; \(u\) kodb is a 0 - 1 variable. If item \(k\), \(k\in M\) in order \(o\in O\) is picked in batch \(b\in B\) at sorting station \(d\) d then \(u\) kodb \( = 1\), otherwise \(u\) kodb \( = 0\); \(y\) ijdb is a continuous variable representing the load on arc \((i, j)\); In the order batching and picking path planning model, process each batch through sub-models (8) and (9), where: if item k is picked from shelf s, then this shelf must be visited; represent the formation of a line graph and form a loop through sub-model (10). Among them, sub-models (8) to (10) include: Among them, x ijdb represents a variable with a value between 0 and 1, and x jidb represents another variable with a value between 0 and 1, and x sjdb represents yet another variable with a value between 0 and 1; In batch b at sorting station d, when passing through arc (i, j), x ijdb = 1, otherwise x ijdb = 0; In batch b at sorting station d, when passing through arc (j, i), x jidb = 1, otherwise x jidb = 0; At the sorting station d, in batch b, when passing through arc (s, j), x sjdb = 1, otherwise x sjdb = 0.

2. The method according to claim 1, characterized in that In S1, it includes: Real-time collect the order demand information of the production system within a specified time period. In the order demand information, it includes at least: raw material item number information and demand quantity information.

3. The method according to claim 1, wherein In S2, it includes: Establish an objective function for the corresponding order picking task: Among them, d ij represents the distance between node i and node j, z represents the total distance of the goods required to sort all orders, A represents the set of arcs between any two nodes, d ∈ V D , b ∈ B d , V D represents the set of all sorting stations, B d represents the batch set of each sorting station; Among them, the preset path planning strategy adopts the S-shaped path strategy.

4. The method according to claim 1, characterized in that, In the order batching and picking path planning model, it also includes: sub-models (11) to (15); Constrain that each batch can only form one loop through sub-model (11); Constraining the load starting from the sorting station through the sub-model (12); Constraining that the remaining load before reaching the required shelf to be accessed is greater than or equal to the total weight of the items to be retrieved from that shelf through the sub-model (13); Constraining that the load between the shelves not to be accessed in each batch is 0 through the sub-model (14); Eliminating sub-circuits through the sub-model (15); Among them, the sub-models (11) to (15) include:

Images