Flexible job shop scheduling method and system based on genetic algorithm

Abstract

The invention belongs to the related technical field of flexible job shop scheduling, and discloses a flexible job shop scheduling method and system based on a genetic algorithm. The method comprises the following steps of: establishing a coding vector by using workpiece batching, process sorting, machine assignment and transportation equipment selection information of the flexible job shop; establishing a compliance function and a comprehensive constraint function as constraint conditions based on the sequential arrangement of processes, the selection of machines and the loading capacity and path of transportation equipment; and constructing an initial population by taking the minimum completion time as an objective function and taking the coding vector as an individual, and solving the objective function by adopting a genetic algorithm to obtain an optimal individual corresponding to the minimum completion time so as to realize the scheduling of the flexible job shop. According to the method and the device, the technical problems of difficulty in workpiece batching, complexity in coding, poor decoding quality and low algorithm solving efficiency in batch scheduling of the flexible job shop under limited transportation resources are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field related to flexible job shop scheduling, and more specifically, relates to a flexible job shop scheduling method and system based on a genetic algorithm. Background Art

[0002] In modern manufacturing, flexible job shops have become increasingly important because they can efficiently handle diverse and personalized production requirements. Although flexible job shops offer advantages in production efficiency and flexibility, they also face some challenges, especially in effective job scheduling under limited transportation resources. Limitations of transportation resources, including the number and capacity of transportation equipment, may become bottlenecks restricting production efficiency. In addition, the batching strategy of workpieces also has a significant impact on production efficiency. Reasonable batching can reduce the number of transports and waiting times, thereby shortening the production cycle and reducing costs.

[0003] However, existing flexible job shop scheduling methods mainly focus on flexible job shop scheduling under limited transportation resources and flexible job shop batching scheduling, without combining the two, and without fully considering the impact of transportation equipment capacity limitations and workpiece batching strategies on the scheduling effect. Therefore, developing a new algorithm that can effectively optimize the batching scheduling of workpieces under limited transportation resources to improve production efficiency, reduce the production cycle, and lower costs is an urgent problem to be solved in this field. Summary of the Invention

[0004] In view of the above-mentioned defects or improvement requirements of the prior art, the present invention provides a flexible job shop scheduling method and system based on a genetic algorithm, which solves the technical problems of difficult workpiece batching, complex coding, poor decoding quality, and low algorithm solving efficiency in flexible job shop batching scheduling under limited transportation resources.

[0005] To achieve the above object, according to one aspect of the present invention, there is provided a flexible job shop scheduling method based on a genetic algorithm, the method comprising the following steps:

[0006] Establish a coding vector using workpiece batching, operation sequencing, machine assignment, and transportation equipment selection information of a flexible job shop; based on the sequential arrangement of operations, machine selection, and the load capacity and path of transportation equipment, establish a compliance function and a comprehensive constraint function as constraint conditions; use the minimum makespan as the objective function, construct an initial population with the coding vector as individuals, and use a genetic algorithm to solve the objective function to obtain the optimal individual corresponding to the minimum makespan, thereby realizing the scheduling of the flexible job shop.

[0007] Further preferably, the coding vector is as follows:

[0008] E(i) = [Bsel(i), Oseq(i), Msel(i), Resl(i)]

[0009] Among them, Bsel(i) is the index of the batch to which workpiece i belongs, Oseq(i) is the process execution sequence information of workpiece i, Msel(i) is the machine assignment result of workpiece i for each process, and Rsel(i) is the transportation equipment assignment result of the batch where workpiece i is located in the handling link.

[0010] Further preferably, the formula of the compliance function is as follows:

[0011] Con(E(i)) = δ1(Bsel(i), Cap, w(i)) + δ2(Oseq(i), m i ) + δ3(Msel(i), D)

[0012] + δ4(Rsel(i), Ctrans)

[0013] Among them, δ1(Bsel(i), Cap, w(i)) is used to determine whether the batch assigned to workpiece i meets the capacity limit. The batch capacity function Cap represents the limit of the batch to accommodate workpieces, and w(i) represents the unit volume or weight of workpiece i; δ2(Oseq(i), m i ) is used to detect whether the process sequence of workpiece i is reasonable, and m i represents the set of optional machines for the process of workpiece i; δ3(Msel(i), D) is used to determine whether the machine selection matches the processing time matrix D; δ4(Rsel(i), Ctrans) is used to calculate whether the transportation equipment assignment meets the load and path cost limitations, and Ctrans represents the transportation cost function.

[0014] Further preferably, the formula of the comprehensive constraint function is as follows:

[0015] Crt(i, o, m, r) = α·Zswitch(o r-1 , o) + β·Ztrans(b r-1 , b r , r) + γ·T(i, o, m)

[0016] Among them, Zswitch(o r-1 , o) is the switching time between the current process o and the previous process ; Ztrans(b r-1 , b r , r) is the load delay on the transportation equipment r during batch switching, T(i, o, m) is the basic processing time of process o on machine m, that is, obtained from D(i, o, m) in step S1, and α, β, γ are weight coefficients, representing the influence proportions of switching overhead, transportation delay, and basic processing time in the scheduling evaluation respectively.

[0017] Further preferably, the objective function is as follows:

[0018] minC max = max(C i ),

[0019]

[0020] where C i is the maximum completion time of workpiece i, is the waiting time before the h-th process of the u-th batch of transporting workpiece i, is the startup setup time of the h-th process of the u-th batch of machining workpiece i by the machine, is the processing time of the h-th process of the u-th batch of machining workpiece i by the machine, is the loading time of the h-th process of the u-th batch of AGV transporting workpiece i.

[0021] Further preferably, the formula of the fitness function is as follows:

[0022]

[0023] where Γ is a benchmark constant related to the production target, δ(i, o) is the scheduling penalty term of workpiece i on process o, and O i is the set of processes of workpiece i, O i = {1, 2,..., K i}.

[0024] According to another aspect of the present invention, there is provided a flexible job shop scheduling system based on a genetic algorithm, which includes an actuator for executing the above-mentioned flexible job shop scheduling method based on a genetic algorithm.

[0025] According to another aspect of the present invention, there is provided a computer-readable storage medium on which a computer program is set, and the computer program realizes the above-mentioned flexible job shop scheduling method based on a genetic algorithm when executed by an actuator.

[0026] Generally speaking, compared with the prior art, the above technical solutions conceived by the present invention have the following beneficial effects:

[0027] 1. The present invention solves the technical problems of difficult job batching, complex coding, poor decoding quality, and low algorithm solving efficiency in flexible job shop batching scheduling under limited transportation resources by using the workpiece batching, operation sequencing, machine assignment, and transportation equipment selection information of a flexible workshop to establish a coding vector, and using this coding vector as an individual in the population of the genetic algorithm. By calculating the average fitness value of each sub-population, the worst-performing sub-population is eliminated, and at the same time, the sub-population containing the optimal individual is retained.

[0028] 2. The present invention constructs a multi-layer coding model of "batching - operation - machine - transportation" for a flexible job shop, decouples the batch constraint, transportation capacity limit, and operation dynamic switching characteristics into independent coding layers and establishes a structured mapping. Traditional scheduling methods usually adopt a single coding or fixed batching strategy, which is difficult to cope with the complex solution space search under multi-constraint coupling. Through the collaborative mapping mechanism of multi-layer coding, the dynamic association relationship between batch allocation, operation sequence, machine selection, and transportation equipment can be accurately described, improving the compatibility of the scheduling scheme with the actual constraints of the workshop and avoiding the problem of local convergence of the solution space caused by the lack of coding dimensions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a flowchart of a flexible job shop scheduling method based on a genetic algorithm constructed according to a preferred embodiment of the present invention;

[0030] Figure 2 is a schematic diagram of the construction of a coding vector constructed according to a preferred embodiment of the present invention;

[0031] Figure 3 is a schematic diagram of population update in the genetic algorithm constructed according to a preferred embodiment of the present invention;

[0032] Figure 4 is a Gantt chart corresponding to the minimized completion time obtained by solving according to a preferred embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0033] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0034] As Figure 1 shown, a workshop scheduling optimization method based on multi-layer coding evolution scheduling and conditional diffusion includes the following steps:

[0035] S1. Obtain basic data such as workpieces, processes, machines, and transportation equipment, and construct a processing time matrix, batch capacity, and transportation cost model; map the information of workpiece batching, process sequencing, machine assignment, and transportation equipment selection into a four-layer coding vector, and use a compliance function to ensure that the coding meets batch and transportation constraints.

[0036] (1) Obtain the set of workpieces J. For each workpiece i, determine the number of processes K i , and establish the set of processes O i = {1, 2,..., K i}}, and the total set of processes formed by all workpieces is denoted as Ω = ∪(i = 1 to N)(O i ).

[0037] Obtain the set of machines M = {1, 2,..., Mt}, where Mt is the number of available machines in the workshop;

[0038] Determine the processing time parameter T(i, o, m) between the workpiece type and machine capacity, which is the standard processing time required for the process o of workpiece i to be executed on machine m.

[0039] Introduce the processing time matrix D, whose dimension is N × (max(K i )) × Mt, denoted as:

[0040] D(i, o, m) = T(i, o, m)

[0041] where D(i, o, m) is the position element of matrix D in the i-th row, o-th column, and m-th layer, i ∈ [1, N], o ∈ [1, K i , m ∈ [1, Mt]. It represents the processing time of the o-th process of the i-th workpiece on the m-th machine.

[0042] (2) According to the maximum batch limit Qb, define the batch set B = {1, 2,..., Bt}, where Bt is the upper limit of the number of batches that can accommodate all workpieces. For each batch b ∈ B, establish a batch capacity function Cap(b), which represents the limit of the number of workpieces accommodated in batch b; define the set of transportation equipment R = {1, 2,..., Rt}, where Rt is the number of all schedulable transportation tools. For each transportation equipment r ∈ R, record its load capacity and feasible path information. Further establish a transportation cost function Ctrans(b, r), which represents the unit distance cost when moving batch b on transportation equipment r. Denote the transportation path length as L(b, r), so as to obtain the total transportation cost as Ctrans(b, r) × L(b, r); further, to quantitatively describe the batch load, this application proposes to construct a batch load determination function Fb(i, b):

[0043] Fb(i, b) = 1 when workpiece i is assigned to batch b;

[0044] Fb(i, b) = 0 when workpiece i is not in batch b.

[0045] Then each batch b has a capacity limit:

[0046] ∑(i = 1 to N)[w(i)·Fb(i, b)] ≤ Cap(b)

[0047] where w(i) represents the unit volume or weight of workpiece i. Through the constraints of Fb(i, b) and Cap(b), it can ensure that the batch load does not exceed the set limit during multi-layer coding.

[0048] (3) As Figure 2 shown, define the multi-layer coding vector E(i) = [Bsel(i), Oseq(i), Msel(i), Rsel(i)], where Bsel(i) is the index of the batch to which workpiece i belongs, Oseq(i) is the process execution sequence information of workpiece i, Msel(i) is the machine assignment result of workpiece i for each process, and Rsel(i) is the transportation equipment assignment result of the batch where workpiece i is located during the handling process; perform an overall mapping on E(i) to generate Etotal = [E(1), E(2),..., E(N)], and Etotal is the set of total vectors for multi-layer coding.

[0049] S2 Construct the compliance function and the comprehensive constraint function

[0050] (1) Uniformly describe the association between multi-layer coding among workpieces, batches, processes, machines, and transportation equipment, and define the compliance function Con(E(i)):

[0051] Con(E(i)) = δ1(Bsel(i), Cap, w(i) + δ2(Oseq(i), m) + δ3(Msel(i), D)

[0052] + δ4(Rsel(i), Ctrans)

[0053] where δ1(Bsel(i), Cap, w(i)) is used to determine whether the batch assigned to workpiece i meets the capacity limit, δ2(Oseq(i), m i ) is used to detect whether the process sequence of workpiece i is reasonable, δ3(Msel(i), D) is used to determine whether the machine selection matches the processing time matrix D, and δ4(Rsel(i), Ctrans) is used to calculate whether the transportation equipment assignment meets the load capacity and path cost limit;

[0054] When Con(E(i)) returns a value of 0, it indicates that this part of the multi-layer coding meets the workshop feasibility constraints; when Con(E(i)) returns a value greater than 0, it means that the coding does not meet the corresponding constraints, and subsequent corrections are made through improved evolutionary operations.

[0055] (2) Construct a comprehensive constraint function that combines the operation switching time and transportation load;

[0056] Based on the scenario characteristics of the comprehensive batch limit and transportation capacity constraints, construct a comprehensive constraint function that includes the operation switching cost and transportation load cost.

[0057] Introduce the operation switching time function Zswitch(o1, o2), which represents the setup time required when switching from operation o1 to operation o2. o1 and o2 belong to the set of operations Ω in step S1.

[0058] Define the transportation load determination function Ztrans(b1, b2, r), which represents the additional waiting or loading time required when relaying batches b1 and b2 on transportation equipment r.

[0059] Incorporate the operation switching time and transportation load cost into the comprehensive constraint function Crt(i, o, m, r) in a weighted manner. This function is used to evaluate the real-time comprehensive cost of workpiece i when selecting machine m and transportation equipment r for operation o.

[0060] The comprehensive constraint function Crt(i, o, m, r) is specifically as follows:

[0061] Crt(i, o, m, r) = α · Zswitch(o r-1 , o) + β · Ztrans(b r-1 , b r , r) + γ · T(i, o, m)

[0062] Among them, Zswitch(o r-1 , o) is the switching time between the current operation o and the previous operation . Ztrans(b r-1 , b r , r) is the load delay on transportation equipment r during batch switching. T(i, o, m) is the basic processing time of operation o on machine m, that is, obtained from D(i, o, m) in step S1. α, β, and γ are weight coefficients, representing the influence ratios of switching cost, transportation delay, and basic processing time in scheduling evaluation respectively. o r-1 , b r-1 respectively represent the operation and batch of the current workpiece i in the previous stage.

[0063] Through the calculation of Crt(i, o, m, r), the feasibility and cost of each layer of coding decision can be comprehensively evaluated according to the real-time production requirements.

[0064] S3 uses a genetic algorithm to solve

[0065] First, parameter settings are carried out, such as population size, number of subpopulations, number of iterations, crossover probability, mutation probability, elitism ratio, etc. Subsequently, population initialization is performed to generate several subpopulations. Each subpopulation randomly generates a batching scheme, and all individuals in each subpopulation are randomly initialized and the fitness evaluation of the individuals is completed simultaneously. The initialized population is used as the parent generation to perform crossover and mutation within each subpopulation to generate offspring, and the fitness evaluation of the offspring is carried out. The parent and offspring are combined, and selection is performed to generate new offspring. Some excellent individuals are selected from the new offspring for local search. If the current number of iterations is a multiple of 10, subpopulations are eliminated and new subpopulations are randomly initialized and generated. Otherwise, it is judged whether the algorithm termination condition is satisfied. If satisfied, the algorithm operation is terminated; if not satisfied, the number of iterations is incremented by one and the iteration continues until the termination condition is met. The termination condition of the algorithm is to reach the maximum number of iterations or the optimal solution has not been updated for 20 consecutive generations.

[0066] (1) As Figure 3 shown, an initial population is generated using crossover and adaptive mutation operators (dynamically adjusted according to transportation and switching differences), and unqualified codes are quickly eliminated.

[0067] In the Bsel(i) and Rsel(i) parts of the multi-layer coding vector Etotal = [E(1), E(2),..., E(N)], according to the batch capacity function Cap(b) and the load capacity of transportation equipment, each workpiece i is assigned to a certain batch in a random or load-balancing strategy-based manner, and transportation equipment that meets the load conditions is selected. In the Oseq(i) and Msel(i) parts, the operation execution order and machine coding are randomly selected, and the comprehensive constraint function Crt is used for quick determination. If there is a situation of overloading or abnormal switching time, the random assignment is readjusted.

[0068] Through multiple repeated constructions and corrections, the initial population is output:

[0069] P0 = {Etotal 1 , Etotal 2 ,..., Etotal P}, where P is the initial population size.

[0070] Furthermore, to accelerate the generation process of the initial population, a construction of an initial feasibility determination function ψE(i)) is proposed:

[0071] ψ(E(i)) = 1 / [1 + ∑(x ∈ Z(i)) Crt(i, ox , m x , r x )]

[0072] Among them, Z(i) represents the set of indices of the process, machine, and transportation batch combinations associated with E(i), and Crt(i, o x , m x , r x ) represents the comprehensive constraint function value of workpiece i under index x. The larger ψ(E(i)) is, the higher the feasibility of the corresponding code E(i). When ψ(E(i)) is too low, local adjustment or re-random allocation is performed.

[0073] (2) Objective function

[0074] The optimization objective is to minimize the makespan C max , as follows:

[0075] min C max = max(C i )

[0076]

[0077] Among them, C i is the makespan of workpiece i, is the waiting time before the h-th process of the u-th batch of transporting workpiece i, is the start-up setup time for the machine to process the h-th process of the u-th batch of workpiece i, is the processing time for the machine to process the h-th process of the u-th batch of workpiece i, is the loading time for the AGV to transport the h-th process of the u-th batch of workpiece i.

[0078] (3) Fitness function

[0079] A. For the coding vector ECtotal in the Bsel(i) or Rsel(i) layer: If it is detected that the current batch b has too high a production capacity utilization rate or the transportation equipment r will have a load conflict in subsequent batches, then the batch or transportation equipment is reallocated and mutated according to the random rate pb.

[0080] For the coding vector ECtotal in the Oseq(i) or Msel(i) layer: If the process switching time Zswitch fluctuates significantly, then the exchange mutation of the process sequence or machine selection is performed according to the random rate po to re-match in the scheduling scenario

[0081] During the adaptive mutation process, set the mutation amplitude coefficient λ, and gradually adjust pb and po according to the current generation and population diversity, so that the mutation is larger in the early iteration and gradually converges in the later iteration.

[0082] Then the adaptive mutation probability calculation function Vadapt(g) is

[0083] Vadapt(g) = λ·(1 - Γ(g) / Γmax)

[0084] where g is the current genetic iteration generation, Γ(g) is the diversity measure of the g-th generation population, measured based on variance or information entropy, Γmax is the maximum diversity observable in the initial population or historical population, and λ is the benchmark coefficient for regulating the overall mutation amplitude.

[0085] B. Extract the population P0 = {Etotal 1 , Etotal 2 ,..., Etotal P} in the initial and early evolution stages from the output multi-layer coding evolutionary scheduling, where P represents the population size.

[0086] For each individual Etotal x (x = 1, 2,..., P), obtain its fitness F(Etotal x ), and uniformly record it in the fitness list Flist = {F(Etotal 1 ), F(Etotal 2 ),..., F(Etotal P )}.

[0087] According to the comprehensive constraint function Crt(i, o, m, r) and the batch and machine selection constraint situations, correspond the information of "workpiece batching scheme, available machine assignment, transportation equipment assignment" involved in the individual with the fitness, and store it in the initial solution database.

[0088] The fitness evaluation function F(Etotal) of the multi-layer coding individual is

[0089]

[0090] where Γ is the benchmark constant related to the production target, used to normalize the scheduling target, and δ(i, o) is the scheduling penalty term of workpiece i in operation o, which is the sum of the comprehensive costs generated by calculating the machine time, transportation delay or switching overhead according to Crt(i, o, m, r).

[0091] When F(Etotal) is close to Γ, it indicates that the comprehensive scheduling performance of this coding solution is relatively good.

[0092] (3) Search algorithm

[0093] After the individuals in the sub-population complete the crossover and mutation operations, fitness evaluation and sorting are performed, and local search is carried out on the dominant individuals. Six neighborhood structures implemented by operating chromosome coding are designed, and some neighborhood structures will generate a certain number of infeasible solutions. These infeasible solutions are removed through chromosome inspection. Chromosome inspection determines the feasibility of the coding by traversing the coding of the new individual and checking whether the MS layer and the OS correspond one by one. The following is an introduction to the neighborhood structures:

[0094] N1: Randomly select a process of a sub-batch to change the processing machine (there are multiple optional processing machines)

[0095] N2: Randomly select a process of a sub-batch to change the transportation equipment (there are multiple transportation equipments)

[0096] N3: At any position in the TS sequence, exchange the transportation equipment for the transportation tasks of subsequent sub-batch processes

[0097] N4: Randomly exchange two adjacent transportation tasks on the same transportation equipment

[0098] N5: Randomly exchange adjacent processing tasks on the same machine

[0099] N6: At any position in the MS sequence, select a group of parallel machines and exchange the subsequent processing tasks.

[0100] To improve the search efficiency of the algorithm, only 20% of the individuals in the sub-population (randomly selected from the top 40% of the individuals in the sub-population in terms of fitness) are selected for local search. The selected individuals perform local search using all neighborhood structures in a certain order. If the fitness value of the newly generated individual is better than that of the original individual, replacement is carried out. The number of local search iterations for each individual does not exceed 50 times.

[0101] From the candidate solutions corrected through multiple iterations, sort according to the comprehensive objective function (including processing duration, transportation delay, switching cost), and select the best-performing solution; after decoding, output the specific batch, process, machine, and transportation plan to achieve real-time flexible job shop scheduling.

[0102] The present invention will be further described below with reference to specific embodiments.

[0103] This embodiment provides a flexible job shop high-efficiency scheduling system based on a multi-population strategy and an equal quantity consistent batching strategy.

[0104] As shown in Table 3, it is a production order of a certain factory. The first column shows the workpiece number and the processing quantity of the workpiece. Each subsequent column is the workpiece process number. For example, the first process of workpiece 1, M3; M4(6 / 6) means that this process can be processed on machines 3 and 4, and the processing time is 6 for both.

[0105] Table 3 Production Order

[0106]

[0107] The factory workshop includes loading and unloading stations, logistics channels, and production and processing equipment; the logistics transportation time between each equipment needs to be considered; the transportation time is affected by the AGV's restriction ability.

[0108] Use the proposed algorithm to schedule the production of this order and obtain the four-layer coding of an excellent scheduling plan for this order, as follows:

[0109] BS: 1,1,1,2,1

[0110] OS: 2,5,1,1,6,3,2,1,8,2,4,7,5,2,3,7,8,1,6,4,3,8,5,2,0,6,0,5,0,4,0,0,7,0,0,0,0,0,0

[0111] MS: 8,11,7,6,10,2,4,9,11,3,6,10,8,5,7,2,6,10,4,8,1,7,9,11,6,3,2,5,7,9,8,4,1,3,10,5,11,1,9

[0112] TS: 1,2,1,1,2,2,1,2,1,1,2,2,1,2,1,2,1,1,1,1,1,2,1,1,1,1,2,2,2,1,1,1,1,1,1,2

[0113] Decode this string of codes to obtain a complete scheduling plan. The batching plans for the 5 types of workpieces are as follows: Workpiece 1 (6, 7); Workpiece 2 (7, 7, 7); Workpiece 3 (7, 7); Workpiece 4 (9, 9); Workpiece 5 (8, 8). Plot the decoding results into a Gantt chart, as Figure 4 shown. The completion time of the order is 270. In the figure, M1 to M8 are processing machines, T1 and T2 are the AGVs for transportation, the blank blocks are the empty transportation trips of the AGVs, and the two identical blocks before and after on the machine respectively represent the setup time and processing time for processing this sub-batch. If the two consecutive sub-batches are of the same type of workpiece, the machine does not need to be switched and there is no setup time.

[0114] Those skilled in the art can easily understand that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A flexible job shop scheduling method based on genetic algorithm, characterized in that The method includes the following steps: Establish a coding vector by using the information of job batching, operation sequencing, machine assignment, and transportation equipment selection in a flexible job shop; based on the sequential arrangement of operations, the selection of machines, and the load and path of transportation equipment, establish a compliance function and a comprehensive constraint function as constraint conditions; take minimizing the makespan as the objective function, construct an initial population with the coding vector as individuals, and use a genetic algorithm to solve the objective function to obtain the optimal individual corresponding to the minimized makespan, so as to realize the scheduling of the flexible job shop.

2. The flexible job shop scheduling method based on genetic algorithm according to claim 1, characterized in that The coding vector is as follows: E(i) = [Bsel(i), Oseq(i), Msel(i), Rsel(i)] Wherein, Bsel(i) is the index of the batch to which workpiece i belongs, Oseq(i) is the operation execution sequence information of workpiece i, Msel(i) is the machine assignment result of workpiece i on each operation, and Rsel(i) is the transportation equipment assignment result of the batch where workpiece i is located in the handling link.

3. The flexible job shop scheduling method based on genetic algorithm according to claim 2, wherein The formula of the compliance function is as follows: Con(E(i)) = δ1(Bsel(i), Cap, w(i)) + δ2(Oseq(i)m i ) + δ3(Msel(i), D) + 84(Rsel(i), Ctrams) Among them, δ1(Bsel(i), Cap, w(i)) is used to determine whether the batch assigned to workpiece i meets the capacity limit. The batch capacity function Cap represents the limit of the batch to accommodate workpieces, and w(i) represents the unit volume or weight of workpiece i; δ2(Oseq(i), m i ) is used to detect whether the operation sequence of workpiece i is reasonable, and m i represents the set of optional machines for the operations of workpiece i; δ3(Msel(i), D) is used to determine whether the machine selection matches the processing time matrix D; δ4(Rsel(i), Ctrans) is used to calculate whether the transportation equipment assignment meets the load capacity and path cost limits, and Ctrans represents the transportation cost function.

4. A flexible job shop scheduling method based on genetic algorithm according to claim 1 or 2, characterized in that The formula of the comprehensive constraint function is as follows: Crt(i, o, m, r) = α·Zswitch(o r-1 , o) + β·Ztrans(b r-1 , b r , r) + γ·T(i, o, m) where Zswitch(o r-1 , o) is the switching time between the current process o and the previous process o r-1 , Ztrans(b r-1 , b r , r) is the load delay on the transportation equipment r during batch switching, T(i, o, m) is the basic processing time of process o on machine m, and α, β, γ are weight coefficients, representing the influence proportions of switching overhead, transportation delay, and basic processing time in scheduling evaluation respectively.

5. A flexible job shop scheduling method based on genetic algorithm according to claim 1 or 2, characterized in that, The objective function is as follows: minC max = max(C i ) Among them, C i is the makespan of workpiece i, is the waiting time before the h-th process of the u-th batch of transporting workpiece i, is the start-up setup time for the machine to process the h-th process of the u-th batch of workpiece i, is the processing time for the machine to process the h-th process of the u-th batch of workpiece i, is the loading time for the AGV to transport the h-th process of the u-th batch of workpiece i.

6. A flexible job shop scheduling method based on genetic algorithm according to claim 1 or 2, characterized in that The formula of the fitness function is as follows: Among them, Γ is a benchmark constant related to the production target, δ(i, o) is the scheduling penalty term of workpiece i in process o, and O i is the set of processes O i ={1, 2,..., K i}.

7. A flexible job shop scheduling system based on genetic algorithm, characterized in that, The system includes an actuator, and the actuator is used to execute a flexible job shop scheduling method according to any one of claims 1-6 based on a genetic algorithm.

8. A computer-readable storage medium having a computer program set thereon, characterized in that, When the computer program is executed by the actuator, it realizes a flexible job shop scheduling method according to any one of claims 1-6 based on a genetic algorithm.

Images