A distributed job shop scheduling method for additive and subtractive hybrid manufacturing

Abstract

The present application relates to the field of intelligent manufacturing, and more particularly to a distributed workshop scheduling method for additive and subtractive hybrid manufacturing. The present application provides a two-stage algorithm, including a batching stage and a scheduling stage. In the batching stage, a balanced height distribution greedy algorithm with two-dimensional interval constraints and space block left-down placement strategy is designed to batch all the workpieces to be produced, determine the workpiece batches, and take the obtained batches as the coding information of the scheduling stage algorithm; in the scheduling stage, a non-dominated sorting genetic algorithm-II with excellent degree strategy driven variable neighborhood search is designed to perform additive and subtractive hybrid manufacturing scheduling, so as to achieve the optimization goal of minimizing the maximum completion time and the total production cost. The present application can realize additive and subtractive hybrid manufacturing scheduling, better utilize resources, maximize production efficiency, and reduce production cost through collaborative work of multiple workshops and load dispersion.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent manufacturing, in particular to a distributed workshop scheduling method for additive and subtractive hybrid manufacturing. BACKGROUND

[0002] Additive manufacturing, also known as 3D printing, is a manufacturing technology that builds three-dimensional entities by layering materials. Additive manufacturing has the advantages of high processing flexibility, high material utilization, and the ability to manufacture complex structures. However, additive manufacturing also has some limitations, such as lower manufacturing precision and strength than traditional manufacturing methods.

[0003] Subtractive manufacturing is a manufacturing technology that obtains the required part shape by removing excess material. Common processes include turning and milling. Subtractive manufacturing technology is mature and widely used. However, subtractive manufacturing has certain limitations in handling complex geometries and internal structures, and the material utilization is relatively low.

[0004] Based on the limitations of additive manufacturing and subtractive manufacturing, additive and subtractive hybrid manufacturing has gradually developed and been applied. Additive and subtractive hybrid manufacturing combines the advantages of additive and subtractive processes, meeting the demand for complex structures while improving processing precision through subtractive processes. Therefore, additive and subtractive hybrid manufacturing has shown broad application prospects in high-precision part and functionally complex part manufacturing.

[0005] Although additive and subtractive hybrid manufacturing solves the problem of complex structure manufacturing and precision machining, it still faces challenges in production scheduling technology in practical applications. Additive and subtractive hybrid manufacturing involves the combination of additive and subtractive processes, and different equipment is required between different processes. Existing production scheduling methods cannot meet the complex scheduling needs of multiple processes and multiple resources. To improve the efficiency of additive and subtractive hybrid manufacturing, an intelligent scheduling algorithm is urgently needed to solve the production scheduling problem of additive and subtractive hybrid manufacturing. SUMMARY

[0006] To further improve the manufacturing efficiency and precision of intelligent manufacturing, and to address the problem that existing production scheduling methods cannot meet the complex scheduling needs of multiple processes and multiple resources, the present application provides a distributed workshop scheduling method for additive and subtractive hybrid manufacturing. The method is a two-stage algorithm, including two stages. The first stage is the batching stage, which is used to obtain the batching result of the workpieces to be produced as the coding information for the second stage algorithm. The second stage is the scheduling stage, which is used to schedule the production of additive and subtractive hybrid manufacturing in a distributed workshop.

[0007] (1) In the batch phase, a balanced height distribution greedy algorithm with two-dimensional interval constraints and space block left-down placement strategy is designed to batch all workpieces to be produced and processed, and determine the workpiece batches, which are taken as the coding information of the algorithm in step (2);

[0008] (2) In the scheduling phase, a non-dominated sorting genetic algorithm-II with excellent degree strategy driven variable neighborhood search is designed to perform overall additive-subtractive composite manufacturing scheduling, so as to realize the optimization target of minimizing the maximum completion time and total production cost:

[0009] min{C max , TPC}

[0010]

[0011] TPC = ∑∑∑∑PC ijmf

[0012] In the formula, C max , TPC are the maximum completion time and total production cost respectively, is the maximum completion time of the factory f, PC ijmf is the production and processing cost of the workpiece i in the process j on the machine m in the factory f.

[0013] A storage device stores instructions and data for implementing a distributed job shop scheduling method for additive-subtractive composite manufacturing.

[0014] A distributed job shop scheduling device for additive-subtractive composite manufacturing comprises a processor and a storage device; the processor loads and executes instructions and data in the storage device to implement a distributed job shop scheduling method for additive-subtractive composite manufacturing.

[0015] The technical scheme provided by the application has the beneficial effects that: the application combines the advantages of additive manufacturing and subtractive manufacturing, can quickly manufacture complex-shaped objects during the manufacturing process, and can perform fine control during the manufacturing process, can save material cost, and can realize rapid and simple manufacturing, and can realize high-precision processing. Distributed job shop production distributes production tasks to multiple interrelated job shops or factories for processing. Each job shop has different machine and process processing capabilities, and the production task is composed of multiple processes and needs to be completed on a specific machine. The advantage of the distributed job shop production problem is that resources can be better utilized, production efficiency is maximized, and through the cooperation of multiple job shops, the load is dispersed and the bottleneck is reduced. In addition, through reasonable workpiece allocation and machine scheduling, production cost can be reduced and the task burden of each job shop can be balanced. BRIEF DESCRIPTION OF DRAWINGS

[0016] The present application will be further described below with reference to the accompanying drawings and examples, in which:

[0017] Figure 1 is a schematic diagram of two-stage processing of multiple workshops of additive manufacturing and subtractive machining in an embodiment of the present application.

[0018] Figure 2 is a schematic diagram of placement completion of a batch of workpieces in an embodiment of the present application.

[0019] Figure 3 is a schematic diagram of five-layer coding designed in an embodiment of the present application.

[0020] Figure 4 is a schematic diagram of a batch sequence string in an embodiment of the present application.

[0021] Figure 5 is a schematic diagram of a batch machine selection string in an embodiment of the present application.

[0022] Figure 6 is a schematic diagram of a process sequence string in an embodiment of the present application.

[0023] Figure 7 is a schematic diagram of a process machine selection string in an embodiment of the present application.

[0024] Figure 8 is a schematic diagram of a factory selection string in an embodiment of the present application.

[0025] Figure 9 is a schematic diagram of the working principle of a crossover operator in an embodiment of the present application.

[0026] Figure 10 is a schematic diagram of the working principle of a mutation operator in an embodiment of the present application.

[0027] Figure 11 is a Pareto front display diagram in an embodiment of the present application.

[0028] Figure 12 is a part batching diagram in an embodiment of the present application.

[0029] Figure 13 is a production Gantt chart in an embodiment of the present application.

[0030] Figure 14 is a schematic diagram of the working of a hardware device in an embodiment of the present application. DETAILED DESCRIPTION

[0031] In order to have a clearer understanding of the technical features, objectives and effects of the present application, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0032] Example 1

[0033] For additive and subtractive hybrid manufacturing distributed scheduling problem, the whole production process of the application includes two workshops, and the production process of each workshop includes additive manufacturing and subtractive processing two parts, as shown in Figure 1 The first part is additive manufacturing, which is used for building parts with complex geometry, forming the matrix or blank, and the second part is subtractive processing, which is used for subsequent fine processing.

[0034] The two-stage hybrid algorithm (Two-stage Hybrid Algorithm with EHAG-2DS and NSGA-II-QSDVN for Additive-Subtractive Hybrid Manufacturing Distributed Scheduling) based on EHAG-2DS and NSGA-II-QSDVN of the application divides the whole production optimization problem into two stages to solve: the first stage is the batching stage, and the second stage is the scheduling stage.

[0035] (1) In the batching stage, an equal height distribution greedy algorithm with two-dimensional interval constraint and space block left lower placement strategy is designed to batch all workpieces, determine the workpiece batches, and obtain the batches as the coding information of the algorithm in stage (2);

[0036] (2) In the scheduling stage, a non-dominated sorting genetic algorithm-II with excellent degree strategy driven variable neighborhood search is designed to perform overall production scheduling, so as to realize the optimization objectives of minimizing the maximum completion time and total production cost:

[0037] min{C max , TPC}

[0038]

[0039] TPC=∑∑∑∑PC ijmf

[0040] In the formula, C max , TPC are the maximum completion time and total production cost, is the maximum completion time of the factory f, and PC ijmf is the cost of the workpiece i in the process j produced by the machine m in the factory f.

[0041] The first stage:

[0042] The additive manufacturing process of the present application is selective laser melting (SLM). In SLM, a specified part is formed from a material powder, which is melted layer by layer by a laser beam. The blade spreads a layer of material powder on the building platform of the machine, moves the laser to scan the specified area to melt the powder, and then forms a layer of the part. Cycle after cycle, the part can eventually be fully formed.

[0043] The placement of the workpieces needs to consider three constraint conditions 1)-3):

[0044] 1) The height of the workpiece must not exceed the maximum height that can be accommodated inside the machine

[0045] h k ≤H mf (1)

[0046] where h k and H mf are the maximum height of the workpiece in the k batch and the maximum height that can be accommodated by the m machine of the f factory, respectively.

[0047] If the height of the workpiece exceeds the maximum height that can be accommodated inside the machine, the workpiece will not be able to be built successfully.

[0048] 2) The shape of the building platform space needs to be considered during the placement of the workpiece, not just the area, so the placement of the workpiece must not exceed the building platform:

[0049]

[0050] where (x, y) represents the left lower corner coordinates of the workpiece on the platform, x and y are the left lower corner coordinates of the workpiece on the platform, l and w are the length and width of the workpiece, d is the placement interval of the workpiece, L mf , W mf are the length and width of the casting platform of the m machine of the f factory.

[0051] 3) Irregular surface features and adhesion may occur after the metal powder is melted and cooled, so there must be an interval between each workpiece, and the interval between the placement of each workpiece is limited to:

[0052]

[0053] where (x, y) represents the left lower corner coordinates of the workpiece i on the platform, (x', y') represents the left lower corner coordinates of the workpiece i' on the platform. l and w represent the length and width of the workpiece i, l' and w' represent the length and width of the workpiece i', and formula (3) is the limit condition to be met when the workpiece i is on the left or top of the workpiece i'.

[0054] 4) Batch production time

[0055] t kmf = SetTmf +ScanT mf ·∑(υ+sυ)+RT mf ·h k (4)

[0056] where SetT mf , ScanT mf , RT mf are the batch setup time, the unit volume scan time and the unit height powder spreading time of the 3D printing machine respectively, and υ, sυare the volume of the workpiece and the volume of the support material used in the manufacturing process of the workpiece respectively.

[0057] Therefore, an even height allocation greedy algorithm with 2D spacing constraints and a bottom-left placement strategy for space blocks (EHAG-2DS) is designed in stage (1) to batch all workpieces, determine workpiece batches, improve the utilization rate of the building platform, and reduce the number of batches to reduce production time and cost. The specific steps are as follows:

[0058] Step 1.1: Sort the workpieces in descending order according to their heights;

[0059] Step 1.2: If the workpiece list is not empty, initialize a manufacturing batch batch and initialize the placement space S; otherwise, go to step 1.6;

[0060] Step 1.3: If the workpiece list is not empty, traverse the placement space list S to find the leftmost and lowest space s; otherwise, go to step 1.6;

[0061] Step 1.4: Traverse the workpiece list P, if part i can be placed in s, add i to batch, remove part i from list P, and end the traversal; if not, rotate part i by 90 degrees to determine whether it can be placed in s, if yes, add i to batch, remove part i from list P, and end the traversal;

[0062] Step 1.5: Update the space list S, if S is not empty, go to step 1.3, otherwise, go to step 1.2;

[0063] Step 1.6: Output all batches.

[0064] The bottom-left placement strategy for space blocks is:

[0065] A placement space is defined by a data structure [X, Y, L, W], which represents a rectangular space. (X, Y) is the coordinate of the lower left corner of the space, L is the length of the space, and W is the width of the space.

[0066] To determine whether a workpiece can be placed in the placement space, a spacing d is added due to the need for spacing between workpieces. The lower leftmost space s in the space list S is found, and if l+2d<L and w+2d<W, the workpiece can be placed in s. The placement requires the lower left corner to be aligned, i.e. (x, y) coincides with (X, Y).

[0067] Update the space list S: after placement, the space s is divided into a new space with the right boundary and the upper boundary. Find the leftmost lower space s, and after placing the workpiece, use the upper boundary of the workpiece as the lower boundary of the new space, and use the right boundary of the workpiece as the right boundary of the new space. The remaining space of the original s space is a new space.

[0068] If l+2d<L and w+2d<W are not satisfied, rotate the workpiece by 90 degrees, and again determine whether the workpiece can be placed. If it can, place it and update the space list S; if it cannot, skip this workpiece.

[0069] Continue to place the workpiece and perform the above operations until the space is full.

[0070] During the placement process, some small spaces may be generated that cannot be used, i.e. all workpieces cannot be placed, and these spaces need to be discarded. The specific way is to make the lower boundary of this space flush with the lowest lower boundary of the two adjacent spaces on both sides and merge into a new space, and discard the remaining space.

[0071] As shown in Figure 2 , a schematic diagram of a batch of workpiece placement is shown. Continue to place the workpiece and perform the above operations until the space is full (including the small spaces that are no longer used).

[0072] Second stage:

[0073] The present application designs an improved NSGA-II algorithm: NSGA-II with Quality Strategy-Driven Variable Neighborhood Search (NSGA-II-QSDVN). The NSGA-II algorithm is a very effective method for solving multi-objective optimization problems. It is based on genetic algorithm, and through mechanisms such as non-dominated sorting and crowding distance, it can find a set of Pareto optimal solutions. The specific steps of the improved NSGA-II algorithm are as follows:

[0074] Step 2.1: NSGA-II with VNS driven by elitist strategy is used to obtain the Pareto solutions:

[0075] Step 2.1.1: The parameters of the algorithm are initialized, the solution individuals are coded and the population is initialized according to the results obtained in step (1);

[0076] Step 2.1.2: If the current iteration number is less than the maximum iteration number, go to step 2.1.3; otherwise, end the iteration and output the Pareto solution;

[0077] Step 2.1.3: The fast non-dominated sorting and crowding distance calculation are performed on the population;

[0078] Step 2.1.4: The sorted population is divided into three parts: the front part, the middle part and the back part; in this embodiment, the front part is the first 15%, the middle part is the middle 25%, and the back part is the last 60%;

[0079] Step 2.1.5: The first 15% is reserved, the middle 25% is subjected to variable neighborhood search, and the last 60% is subjected to crossover and mutation to obtain offspring;

[0080] Step 2.1.6: The fast non-dominated sorting and crowding distance calculation are performed on the combined population of the parents and the offspring, and the first popsize individuals are selected to form a new population, where popsize represents the population size;

[0081] Step 2.1.7: The iteration number is incremented by one, and the iteration is continued to step 2.1.2 until the maximum iteration number is reached.

[0082] The problem characteristics in this embodiment are described as follows:

[0083] As shown in Figure 1 , when the workpiece is in the additive manufacturing part, according to the characteristics of the additive manufacturing process, the workpieces in the same batch can be processed simultaneously on one machine, i.e. batch processing; while in the subtractive manufacturing part, a machine cannot process multiple workpieces simultaneously, and a machine can only process one workpiece at a time.

[0084] There are five sub-problems: 1) factory selection, selecting a factory for each batch / workpiece; 2) batch machine selection, selecting a machine to process the workpieces in a batch; 3) batch ordering, ordering the processing sequence of the batches when the additive manufacturing batch processing part; 4) process machine selection, selecting a machine to process; 5) process ordering, ordering the processing of the process when the subtractive manufacturing part is processed separately.

[0085] Additive and subtractive composite manufacturing constraints:

[0086]

[0087] In the formula, C ij is the end time of the workpiece i process j; B ij is the start time of the workpiece i process j; AT ij is the workpiece i process j additive manufacturing part processing time, that is, the processing time of the process in the batch; ST ij is the workpiece i process j subtractive machining part processing time.

[0088] Algorithm implementation:

[0089] 1. Coding method: a five-layer coding is designed, as shown in Figure 3 , which represents five sub-problems:

[0090] Batch sequence string (BS)

[0091] Batch machine selection string (BMS)

[0092] Operation sequence string (OS)

[0093] Operation machine selection string (OMS)

[0094] Factory selection string (FS)

[0095] The algorithm coding embodies the batch information, as in Figure 3 , the workpiece can be batched in 4 processes, batch1 and batch2 represent the batching results obtained according to the algorithm in stage (1), that is, a total of two batches are to be produced, then the batch sequence string (BS) only has "1" and "2", and the BS string has a total of 2*4 gene positions, the length is 8. part1, part2, part3 represent workpiece 1, workpiece 2, workpiece 3. In the figure, batch1: part1, part3 indicates that workpiece 1 and workpiece 3 are divided into batch1; batch2: part2 indicates that workpiece 2 is divided into batch2.

[0096] For the process stage that can be batched, the batch is taken as the scheduling unit, and for the process stage that is processed alone, the process of the single workpiece is taken as the scheduling unit.

[0097] As shown in Figure 4 , the batch sequence string represents the process in the order of the batch number appearing in the chromosome, such as from left to right, the first "2" represents the first process of batch 2, and the second "2" represents the second process of batch 2.

[0098] As shown in Figure 5 , the batch machine selection string represents the machine where each batch is produced, with the index number in each batch optional machine set, for example, the batch1 second process optional machine set is {M4, M5, M6}, then the gene "3" corresponding to batch1 O2 represents M6, where M represents a machine.

[0099] As shown in Figure 6 , the process order string represents the process (plus batch processing) in the order of the job number appearing in the chromosome, from left to right, the first "3" represents the fifth process of job 3 (the first four processes are processed by batch processing), and the first "1" represents the fifth process of job 1.

[0100] As shown in Figure 7 , the process machine selection string represents the machine where each job process is produced, with the index number in each process optional machine set, for example, the fifth process of job 1 optional machine set is {M10, M15}, then the gene "2" corresponding to O15 represents M15.

[0101] As shown in Figure 8 , the factory selection string, each batch / job can only be processed in one factory, and the production process does not span factories. The factory number appearing order represents the factory where each batch is produced, and the factory where the batch is located already contains the factory information where the job is located.

[0102] 2. Step 2.2: Decode the solution individual obtained in step 2.1:

[0103] The previous decoding method may cause some unnecessary idle time of the machine, not only causing resource waste and cost increase, but also possibly increasing the completion time. In order to solve these problems, the present application proposes an active decoding method to reduce idle time, as follows:

[0104] Step 2.2.1: Decode the batch string, traverse the batch string, find the machine m of the factory f selected by the batch b process j;

[0105] Step 2.2.2: Traverse the idle time period of machine m;

[0106] Step 2.2.3: Set the start processing time of batch b process j to be equal to the maximum of the completion time of batch b process j-1 and the start time of the idle time period, judge whether the start processing time of batch b process j plus the processing time is less than the end time of the idle time period, if less than go to step 2.2.4, otherwise go to step 2.2.2;

[0107] Step 2.2.4: update the idle time period of machine m and the completion time and production cost of job i in process j, let the completion time of job i in process j be equal to the completion time of job i in process j-1 and the start time of the idle time period; if the decoding of the job string is finished, go to step 2.2.5, otherwise go to step 2.2.1;

[0108] Step 2.2.5: continue decoding the job string, traverse the job string to find the machine m of the factory f selected by process i in process j;

[0109] Step 2.2.6: traverse the idle time period of machine m;

[0110] Step 2.2.7: let the start processing time of job i in process j be equal to the maximum of the completion time of job i in process j-1 and the start time of the idle time period, judge whether the start processing time of job i in process j plus the processing time is less than the end time of the idle time period, if yes, go to step 2.2.8, otherwise go to step 2.2.6;

[0111] Step 2.2.8: update the idle time period of machine m and the completion time and production cost of job i in process j, if the decoding of the job string is finished, go to step 2.2.9, otherwise go to step 2.2.5;

[0112] Step 2.2.9: decoding is finished, output the maximum completion time and the total production cost.

[0113] 3. Evolutionary strategy

[0114] 1) Elite individual reservation: for the population after non-dominant sorting and crowding degree calculation sorting, take the top 15% of individuals for elite reservation, directly enter the next generation.

[0115] 2) Quality Strategy-Driven Variable Neighborhood Search (QSDVNS): take the middle 25% of individuals in the sorted population for neighborhood search, excellent individuals maintain excellent characteristics, improve the local search ability of the population. Five neighborhood search operators are designed, and the operator variable neighborhood search is selected by driving the strategy quality.

[0116] Step 1: let the population size be m, n = 0.

[0117] Step 2: randomly select a neighborhood search operator by roulette mechanism.

[0118] Step 3: neighborhood search the population with the selected operator, if the new individual dominates the original individual, n+1.

[0119] Step 4: After the evolution of population individuals, calculate n / m, update the population and roulette, and let the roulette correspond to the value of the search operator plus n / m.

[0120] 5 Neighborhood search operators are as follows:

[0121] ① Random process minimum processing time machine transfer

[0122] Randomly select several processes to select the machine with the minimum processing time

[0123] ② Random process minimum production cost machine transfer

[0124] Randomly select several processes to select the machine with the minimum production cost

[0125] ③ Random two batch exchange

[0126] Randomly select two batches from the batch sequence of the factory with the longest completion time to exchange positions

[0127] ④ Random two process exchange

[0128] Randomly select two processes from the process sequence of the factory with the longest completion time to exchange positions

[0129] ⑤ Key factory batch transfer

[0130] Select the factory with the longest completion time, and transfer a random batch in it to another factory

[0131] 3) Crossover, mutation: the remaining 60% of individuals are crossed and mutated to improve the global search ability of the population.

[0132] ① Crossover operator:

[0133] The application proposes a crossover operator suitable for flexible job shop scheduling problem coding, a fixed gene fragment self-exchange crossover operator.

[0134] Two sites are randomly selected, and the gene fragments between the two sites of two individuals are exchanged. In the flexible job shop scheduling problem, the coding needs to ensure the correct number of processes, and simply exchanging gene fragments can result in infeasible solutions, so a new way of exchanging is adopted. As shown in the following figure. The gene fragments to be exchanged are between site 3 and site 6, and the gene fragments corresponding to p1 after exchange are "1, 2, 3, 1", and the gene fragments corresponding to p2 after exchange are "2, 1, 3, 2". When the first exchange occurs, the "2" of p1 needs to be exchanged with the "1" of p2, but since the effectiveness of the chromosome needs to be maintained, it is not exchanged with the gene of p2, but is exchanged with its own gene. Find the first gene appearing from left to right in its own gene and exchange it, and the "2" at site 3 of p1 is exchanged with the "1" at site 2, and the rest are exchanged in the same way. p2 is also exchanged in the same way.

[0135] ② Mutation operator:

[0136] The present application proposes a mutation operator suitable for coding of the flexible job shop scheduling problem, a logistic curve dynamic boundary mapping swap mutation operator

[0137] Mutation probability Mr:

[0138] In the formula, t is the current iteration number, and T is the maximum iteration number. As the iteration changes, the mutation rate also changes, Mr0 represents the initial mutation rate, and Mr t represents the mutation rate at the tth generation.

[0139] The mutation probability gradually decreases with the iteration number, and the logistic curve has the characteristics that the decrease is slow before the inflection point and fast after the inflection point, and finally tends to be stable. The parameter a can control the degree of rapidity of the decrease of the mutation probability, the parameter b controls the inflection point from which iteration, and the parameter c controls the degree of decrease of the curve at the inflection point. By selecting appropriate parameters, a high mutation probability is maintained in the early iteration process, which facilitates the generation of new solutions and improves the exploration ability of the algorithm, and in the later iteration process, the algorithm gradually converges, so that the mutation probability rapidly decreases, preventing the destruction of the excellence of the solution.

[0140] The mutation strategy uses a random exchange method: a few genes are randomly selected for random exchange position, and the corresponding genes after the exchange position are placed in the gene site after calculation by formula (6) to ensure the feasibility of the solution. Wherein, m p is the number of machines that the process can select.

[0141]

[0142] In the formula, gene p gene value of p position of solution p gene value of p position of solution after crossover.

[0143] Experimental parameters in the embodiment of the present application

[0144] The 3D printer machine parameter settings in the experiment are shown in Table 1:

[0145] Table 13D printing machine related parameters

[0146]

[0147] The experimental parameter settings are shown in Table 2:

[0148] Table 2 Algorithm parameter settings in the experiment

[0149]

[0150]

[0151] Through the above parameters, the technical scheme of the embodiment of the present application is implemented, and a Pareto frontier display diagram as shown in Figure 11 is obtained, in which the abscissa is the maximum production task maximum completion time, the ordinate is the total production cost, different color curves represent different dominance relations of individuals in the population, and the solution on the curve closest to the lower left corner is the non-dominated solution set in the population, i.e. the Pareto optimal solution set. The workpiece batching diagram obtained is as shown in Figure 12 , in which a total of 9 batches are divided, and at least 2 workpieces are included in each batch. The production Gantt chart obtained is as shown in Figure 13 , in which different colors represent different batches or workpieces, and different colors in a factory represent different jobs of the same batch or workpiece.

[0152] Embodiment 2

[0153] A distributed workshop scheduling device 401 for additive and subtractive composite manufacturing, as shown in Figure 14 , comprises a processor 402 and a storage device 403; the processor 402 loads and executes instructions and data in the storage device 403 to realize the distributed workshop scheduling method for additive and subtractive composite manufacturing.

[0154] Embodiment 3

[0155] A storage device, the storage device stores instructions and data for realizing the distributed workshop scheduling method for additive and subtractive composite manufacturing.

[0156] The above merely describes preferred embodiments of the present application and is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A distributed workshop scheduling method for additive and subtractive manufacturing, characterized in that: The method includes two stages. The first stage is the batching stage, and the second stage is the scheduling stage; (1) In the batching stage, design a balanced height allocation greedy algorithm with two-dimensional interval constraints and the lower-left placement strategy of spatial blocks to batch all the workpieces to be produced and processed, determine the workpiece batches, and use the obtained batches as the coding information of the algorithm in stage (2); The lower-left placement strategy of spatial blocks is as follows: Define the placement space with a data structure [X, Y, L, W], where (X, Y) are the coordinates of the lower-left corner of the space, L is the length of the space, and W is the width of the space; Judge whether the workpiece can be placed in the placement space. Find the most lower-left placement space s in the space list S. If l + 2d < L and w + 2d < W, where l and w represent the length and width of the workpiece, then the workpiece can be placed in s. The placement requirement is that the lower-left corners are aligned, that is, (x, y) coincides with (X, Y); Update the space list S: After placement, divide the placement space s, dividing it by the right boundary and the upper boundary; find the most lower-left space s. After placing the workpiece, use the upper boundary of the workpiece as the lower boundary of the new space and the right boundary of the workpiece as the right boundary of the new space. The remaining space of the original s space is a new space; If l + 2d < L and w + 2d < W are not satisfied, rotate the workpiece 90 degrees, and then judge whether the workpiece can be placed again. If it can, perform the placement and update the space list S; if not, skip this workpiece; Continue to place workpieces and perform the above operations until the space is full; (2) In the scheduling stage, design a non-dominated sorting genetic algorithm-II with excellence strategy-driven variable neighborhood search to perform distributed workshop additive and subtractive hybrid manufacturing scheduling, and achieve the optimization goal of minimizing the makespan and the total production cost: In the formula, These are the maximum completion time and the total production cost, respectively. It is the maximum completion time of factory f. It is the cost of producing workpiece i in process j on machine m in factory f; The specific implementation process is as follows: Step 2.1: Use a non-dominated sorting genetic algorithm-II with excellence strategy-driven variable neighborhood search to find the Pareto solution: Step 2.1.1: Initialize the algorithm parameters, encode the solution individuals according to the results obtained in step (1), and initialize the population; [[ID= ​ ​ ​ ​ ​ ​ ​ ​ Step 2.2.3: Set the start time of batch b process j to the maximum value between the completion time of batch b process j-1 and the start time of the idle period. Determine whether the start time of batch b process j plus the processing time is less than the end time of the idle period. If it is less, go to step 2.2.4; otherwise, go to step 2.2.

2. Step 2.2.4: Update the idle time period of machine m and the completion time and production cost of process j in batch b. Set the completion time of process j of all workpieces in batch b to be equal to the completion time of process j in batch b. If the batch string is decoded, go to step 2.2.5; otherwise, go to step 2.2.

1. Step 2.2.5: Continue decoding the process string, traverse the process string, and find the machine m of factory f selected for process i and process j; Step 2.2.6: Iterate through the idle time periods (idle_time) of machine m; Step 2.2.7: Set the start time of process j of workpiece i to the maximum value between the completion time of process j-1 of workpiece i and the start time of the idle period. Determine whether the start time of process j of workpiece i plus the processing time is less than the end time of the idle period. If it is less, go to step 2.2.8; otherwise, go to step 2.2.

6. Step 2.2.8: Update the idle time period of machine m and the completion time and production cost of workpiece i and process j. If the process string is decoded, proceed to step 2.2.9; otherwise, proceed to step 2.2.

5. Step 2.2.9: Decoding complete, output maximum completion time and total production cost.

2. The distributed workshop scheduling method for additive and subtractive manufacturing as described in claim 1, characterized in that: In step (1), all workpieces to be produced and processed are divided into batches. The specific steps for determining the workpiece batches are as follows: Step 1.1: Sort the workpieces in descending order based on their height; Step 1.2: If the workpiece list is not empty, initialize a processing batch and initialize the placement space S; Otherwise, proceed to step 1.6; Step 1.3: If the workpiece list is not empty, traverse the placement space list S and find the bottom left space s; Otherwise, proceed to step 1.6; Step 1.4: Traverse the workpiece list P. If part i can be placed in s, add i to batch, remove part i from list P, and end the traversal; if not, rotate part i by 90 degrees to determine if it can be placed in s. If it can, add i to batch, remove part i from list P, and end the traversal; where part i represents the i-th workpiece. Step 1.5: Update the space list S. If S is not empty, go to step 1.3; otherwise, go to step 1.

2. Step 1.6: Output all batches.

3. The distributed workshop scheduling method for additive and subtractive manufacturing as described in claim 2, characterized in that: In step 1.4, the placement of the workpiece needs to satisfy three constraints: 1) The height of the workpiece must not exceed the maximum height that the machine's interior can accommodate. The following restrictions must be met: (1) In the formula, and These are the maximum height of the workpiece in batch k and the maximum height that machine m in factory f can accommodate, respectively. 2) The placement of the workpiece must not exceed the construction platform, and the following restrictions must be met: (2) In the formula, This represents the coordinates of the lower left corner of the workpiece on the platform. l,w These are the length and width of the workpiece. d It refers to the spacing between workpieces. These are the length and width of the casting platform of machine m in factory f; 3) The spacing between each workpiece must meet the following constraints: (3) In the formula, and Representing workpiece i and workpiece i respectively The coordinates of the bottom left corner of the platform, l,w Indicates the length and width of workpiece i. workpiece Length and width.

4. A distributed workshop scheduling method for additive and subtractive manufacturing as described in claim 3, characterized in that: The production time for each batch of workpieces is: (4) In the formula, These are the batch setting time, unit volume scanning time, and unit height powder spreading time for the 3D printer. These refer to the volume of the workpiece within the batch and the volume of the supporting materials used in the workpiece manufacturing process, respectively.

5. A distributed workshop scheduling method for additive and subtractive manufacturing as described in claim 1, characterized in that: Additive manufacturing is a selective laser melting process in which a specified part is formed from material powder, which is melted layer by layer by a laser beam.

6. A distributed workshop scheduling method for additive and subtractive manufacturing as described in claim 1, characterized in that: Constraints of Additive and Subtractive Composite Manufacturing: (5) In the formula, It is the end time of process j for workpiece i; It is the start time of process j for workpiece i; It is the processing time of the additive manufacturing part of workpiece i in process j, that is, the processing time of this process in the batch; It is the processing time of the subtractive processing part of workpiece i in process j.

7. A storage device, characterized in that: The storage device stores instructions and data to implement the distributed workshop scheduling method for additive and subtractive composite manufacturing as described in any one of claims 1 to 6.

8. A distributed workshop scheduling device for additive and subtractive manufacturing, characterized in that: include: A processor and a storage device; the processor loads and executes instructions and data in the storage device to implement the distributed workshop scheduling method for additive and subtractive composite manufacturing as described in any one of claims 1 to 6.

Images