A multi-objective optimization method for energy-saving batch flow scheduling in flexible job shop considering processing preparation resources

Abstract

The present application relates to a kind of flexible job shop energy-saving batch flow scheduling multi-objective optimization method considering processing preparation resources, belong to workshop scheduling technical field.It includes steps: flexible job shop energy-saving batch flow scheduling multi-objective optimization mathematical model considering processing preparation resources;Model solution method based on knowledge-driven batching method and improved multi-objective evolutionary algorithm;Based on the tooling adjustment strategy of adjacent process.The optimization method proposed in the present application can efficiently solve the flexible job shop energy-saving batch flow scheduling multi-objective optimization problem considering processing preparation resources, realize the efficient collaborative optimization of workpiece batching, machine allocation, process sequencing, tooling allocation, and obtain the scheduling scheme with good completion time, total energy consumption and total processing cost.

Description

Technical Field

[0001] This invention relates to a multi-objective optimization method for energy-saving batch flow scheduling in flexible workshops that takes into account processing preparation resources, and belongs to the field of workshop scheduling technology. Background Technology

[0002] In the face of fierce market competition, how to deliver orders efficiently, with high quality, and at low cost while saving energy, thereby enhancing core competitiveness, is a crucial issue for enterprises. Therefore, flexible workshop energy-saving batch flow scheduling, in addition to considering energy consumption targets, also needs to coordinate completion time, processing costs, and other objectives to achieve integrated optimization of economic and green indicators. However, these two types of indicators often have conflicts in their synergistic optimization. Reducing the total energy consumption of the workshop can lead to an extension of workpiece completion time and also have uncertain impacts on workpiece processing costs, posing a challenge to the multi-objective synergistic optimization capabilities of optimization methods. Furthermore, workshop production, while focusing on machine tool and material constraints, also needs to consider the constraints of tooling and fixtures, etc., which play a key role on the production line, ensuring accurate workpiece positioning and stable fixation, directly affecting the smooth operation of the entire production line. At the same time, the allocation of processing preparation resources has a significant impact on production efficiency and resource utilization. Reasonable arrangement and utilization of tooling and fixtures can significantly reduce equipment changeover and preparation time, improving equipment utilization. Therefore, research on multi-objective optimization methods for energy-saving batch flow scheduling in flexible workshops that take into account processing preparation resources is crucial for improving machine tool utilization, reducing tooling and fixture switching times, and lowering waiting time and processing preparation energy consumption.

[0003] The energy-saving batch flow scheduling problem in flexible workshops, considering processing preparation resources, involves multiple factors such as sub-problem coupling, objective conflict, and resource constraint integration, resulting in high computational complexity and numerous local optima. In this context, the matching result between the optimization method and the problem characteristics directly affects the quality of the scheduling solution. Therefore, when modeling energy-saving batch flow scheduling in flexible workshops considering processing preparation resources, it is crucial to comprehensively consider sub-problems such as workpiece batching, process sequencing, machine tool and tooling allocation, and the coordinated optimization of economic and green indicators, as well as processing and processing preparation resources. During the solution process, attention should be paid to the synergistic relationship between problem characteristics, algorithm operations, and algorithm parameters. Analyzing their importance and applicability to algorithm performance is key to obtaining a superior scheduling scheme for the energy-saving batch flow scheduling problem in flexible workshops considering processing preparation resources, and this still requires further systematic and in-depth exploration and research.

[0004] Current research on energy-saving batch flow scheduling in flexible workshops, considering processing preparation resources, largely revolves around traditional economic indicators. It fails to explore the multi-objective collaborative optimization of energy consumption, completion time, and processing costs, or the coupling mechanism of multiple sub-problems in the energy-saving batch flow scheduling problem. Furthermore, existing studies often incorporate processing preparation time into the workpiece processing time, which, while easy to implement, prevents proper optimization of processing preparation resources. Based on the no-free-lunch theory, no single algorithm can be universally applicable to two different optimization problems. Therefore, research on optimization methods for energy-saving batch flow scheduling in flexible workshops, focusing on processing preparation resources, remains crucial. Exploring the collaborative optimization mechanism among sub-problems and investigating the conflict relationships between different scheduling indicators is key to designing efficient multi-objective collaborative optimization methods that ensure the scientific validity and effectiveness of multi-objective decision-making. Summary of the Invention

[0005] To address the aforementioned shortcomings and improvement needs of existing technologies, this invention proposes a multi-objective optimization method for energy-saving batch flow scheduling in flexible workshops that considers processing preparation resources. The aim is to solve the problems of traditional optimization methods being difficult to apply and prone to premature convergence, and to achieve coordinated optimization of processing resources, processing preparation resources, and transportation resources in energy-saving batch flow scheduling of flexible workshops that considers processing preparation resources.

[0006] To address the aforementioned technical problems, this invention proposes a multi-objective optimization method for energy-saving batch flow scheduling in a flexible workshop that considers processing preparation resources. The method includes the following steps: S1, constructing a multi-objective optimization mathematical model for energy-saving batch flow scheduling in a flexible workshop that considers processing preparation resources; S2, proposing a hybrid method combining a knowledge-driven batching method and an improved multi-objective evolutionary algorithm to solve the model; S3, designing a tooling adjustment strategy based on adjacent processes to reduce the number of tooling and fixture changes, shorten workpiece completion time, and reduce processing preparation energy consumption.

[0007] According to the present invention, as a further preferred embodiment, the flexible workshop energy-saving batch flow scheduling multi-objective optimization mathematical model considering processing preparation resources in step S1 has the optimization objectives of minimizing completion time, total energy consumption, and total processing cost, and the calculation formula is as follows:

[0008] F = min(TEC,T) max ,TC)

[0009] Where F is the objective function, TEC, T max TC represents the total energy consumption of the workshop, the completion time, and the total processing cost, respectively.

[0010] In addition to meeting the basic constraints of batch flow scheduling in flexible workshops, it is also necessary to satisfy the constraint that the same tooling can only be applied to one process operation at a time, as expressed as:

[0011]

[0012]

[0013] Among them, ST i,j,r For process Q ij Occupying tooling MT r Start time; RT g,h,r For process Q ij Occupying tooling MT r End time; As a decision variable, if process Q ij With subsequent processes Q gh All require the use of tooling MT r The value is 1 if the condition is met, otherwise it is 0; B is a sufficiently large real number; O i,j For the machinable process Q ij Tooling MT r Set; O g,h For the machinable process Q gh Tooling MT r Set; n' is the total number of workpieces after batch partitioning; i, g are workpiece indices; g, h are process indices; r is tooling index; q i For workpiece J i Number of processes; q g For workpiece J g The number of processes.

[0014] In addition, the constraint that all tooling is available at the initial moment of scheduling must also be satisfied, expressed as:

[0015]

[0016] Y i,j,k,r ∈{0,1},i∈[1,n'],j∈[1,q i ], k∈[1,m], r∈[1,N t ]

[0017] R r ∈{0,1},r∈[1,N t ]

[0018] Where k is the machine tool index; m is the number of machine tools; N t For tooling quantity; BT i,1,k For workpiece J i The first process is on machine tool M k Start time of the upper processing; Y i,j,k,r As a decision variable, if the process Q after batch processing... ij In machine tool M k Using tooling MTr R takes 1 if the condition is met, otherwise takes 0; r As a decision variable, if tooling MT r If available at the initial scheduling time, the value is 1; otherwise, it is 0.

[0019] According to the present invention, as a further preferred embodiment, step S2 includes the following sub-steps:

[0020] A two-stage feedback optimization approach is adopted, utilizing a knowledge-driven batching method framework to optimize the workpiece batching sub-problem, and based on the proposed improved multi-objective evolutionary algorithm, to optimize the machine tool allocation, process sequencing, and tooling allocation sub-problems. Specifically, within the knowledge-driven flexible workshop batching method framework, and considering the three-objective optimization characteristics of the model, an optimization strategy based on the moth-flame operator is introduced to fully explore the three-dimensional solution space of the problem.

[0021] Design a co-evolutionary operator based on genetic and differential evolution operators to promote information exchange among individual moths.

[0022] According to the present invention, as a further preferred embodiment, the tooling adjustment strategy based on adjacent processes described in step S3 includes the following sub-steps:

[0023] Obtain current workpiece, machine tool, and fixture status information;

[0024] Determine the first process to begin among the selected processes. If multiple processes start simultaneously, randomly select one of them.

[0025] Determine if the process is adjacent to the preceding process on the selected machine tool. Select the same tooling for processing. If yes, discard this process and return to step 2; otherwise, proceed to the next step.

[0026] Judgment process Can the tooling be used in the process? Its process Is it available within the processing cycle? If yes, then... Tooling and If the process is consistent, update the tooling status information and proceed to the next step; otherwise, discard the process and return to step 2.

[0027] in, For machine tool M k The j-th process in the upper processing, For machine tool M k The (j-1)th process in the upper processing.

[0028] End the algorithm.

[0029] The beneficial effects of this invention are as follows: It comprehensively considers total workshop energy consumption, workpiece completion time, and total processing cost, constructing a multi-objective optimization mathematical model for flexible workshop energy-saving batch flow scheduling that considers processing preparation resources. A hybrid method combining a knowledge-driven batching approach and an improved multi-objective evolutionary algorithm is proposed to solve the model. Within the framework of the knowledge-driven batching approach, the three-dimensional solution space of the problem is fully explored based on the moth-flame operator. A co-evolutionary operator is introduced to improve population diversity, promote information interaction between populations, and improve the algorithm's search speed and convergence effect. A tooling adjustment strategy based on adjacent processes is designed to reduce the number of tooling and fixture changes, shorten workpiece completion time, and reduce processing preparation energy consumption. The optimal workpiece batching, machine tool allocation, process sequencing, and tooling allocation scheme is obtained, comprehensively considering total workshop energy consumption, workpiece completion time, and total processing cost. This promotes the integration of flexible production batch flow scheduling technology with intelligent optimization, considering multi-resource constraints of processing preparation resources, providing theoretical basis and practical support for multi-variety variable batch production under the collaborative constraints of processing resources and processing preparation resources. Attached Figure Description

[0030] Figure 1(a) and 1(b) Figure 1(a) shows the scheduling scheme before workpiece batching, and Figure 1(b) shows the scheduling scheme after workpiece batching optimization.

[0031] Figure 2 This is a model solution framework diagram of the hybrid method based on the knowledge-driven batch method and the improved multi-objective evolution algorithm of the present invention;

[0032] Figure 3 This is a Gantt chart of the optimized batch scheduling scheme and machine tool allocation for the baseline problem MK06 (Ave_B=20) of this invention;

[0033] Figure 4 This is a Gantt chart showing the optimized batch scheduling scheme, process and tooling allocation for the baseline problem MK06 (Ave_B=20) of this invention.

[0034] Figure 5(a) , 5(b) Figure 5(c) shows the box plots of the comparison results of the knowledge-driven batch method and the improved multi-objective evolutionary algorithm of the present invention with the IGD, GD and MS indices of MOEA / D, NSGA-II, MOPSO, MOEA / PSL and LMOCSO for solving example problems. Figure 5(a) is a box plot of the comparison results of the IGD index, Figure 5(b) is a box plot of the comparison results of the GD index, and Figure 5(c) is a box plot of the comparison results of the MS index.

[0035] Figure 6A Gantt chart illustrating the scheduling scheme for finding the optimal solution to an instance problem using the knowledge-driven batching method and the improved multi-objective evolutionary algorithm, as described in this invention.

[0036] Figure 7 This is a Gantt chart of the current scheduling scheme of a problem enterprise, as exemplified by an embodiment of the present invention. Detailed Implementation

[0037] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. However, it should be understood that the embodiments are used to explain the present invention and are not intended to limit the present invention.

[0038] To make the present invention clearer, the method of the present invention will be described below with reference to a specific embodiment, such as... Figure 1(a) and 1(b) As shown in the figure; the flexible workshop energy-saving batch flow scheduling optimization method provided by the embodiments of the present invention includes the following steps: Figure 2 As shown, it includes the following steps:

[0039] S1. Multi-objective optimization mathematical model for energy-saving batch flow scheduling in flexible workshops, considering processing preparation resources.

[0040] S2. A model solution method based on knowledge-driven batch processing and an improved multi-objective evolutionary algorithm.

[0041] S3. Tooling adjustment strategy based on adjacent processes.

[0042] S4. Numerical Experiment.

[0043] S5, Instance Verification.

[0044] Furthermore, the multi-objective optimization mathematical model for energy-saving batch flow scheduling of flexible workshops, considering processing preparation resources, in step S1 is as follows:

[0045] Objective function: F = min(TEC, T) max ,TC)

[0046] Where F is the objective function, TEC, T max TC represents the total energy consumption of the workshop, the completion time, and the total processing cost, respectively.

[0047] Based on the basic constraints of batch flow scheduling in flexible workshops, the constraints include: treating sub-batch workpieces after batching as approximately one workpiece; when workpieces cannot be completely divided equally, the remaining workpieces are processed in the last batch; the preceding constraint that a workpiece can only be processed after its preceding process is completed; the processing capacity constraint that a machine tool can only process one process at a time; the processing number constraint that each process can only be processed once by any machine tool; the availability of machine tools at the initial scheduling time; and the random arrival constraint of workpieces.

[0048] In addition, the constraint that the same tooling can only be applied to one process operation at a time must be met, expressed as:

[0049]

[0050]

[0051] Among them, ST i,j,r For process Q ij Occupying tooling MT r The start time.

[0052] RT g,h,r For process Q ij Occupying tooling MT r The end time.

[0053] As a decision variable, if process Q ij With subsequent processes Q gh All require the use of tooling MT r If the value is 1, then the value is 0; otherwise, the value is 0.

[0054] B is a sufficiently large real number.

[0055] O i,j For the machinable process Q ij Tooling MT r gather.

[0056] O g,h For the machinable process Q gh Tooling MT r gather.

[0057] n' represents the total number of workpieces after batch division of all workpieces.

[0058] i,g are the workpiece indices;

[0059] g and h are process indices.

[0060] r is the tooling index.

[0061] q i For workpiece J i The number of processes.

[0062] q g For workpiece J g The number of processes.

[0063] In addition, the constraint that all tooling is available at the initial moment of scheduling must also be satisfied, expressed as:

[0064]

[0065] Yi,j,k,r ∈{0,1},i∈[1,n'],j∈[1,q i ], k∈[1,m], r∈[1,N t ]

[0066] R r ∈{0,1},r∈[1,N t ]

[0067] Where k is the machine tool index.

[0068] m represents the number of machine tools.

[0069] N t This refers to the quantity of tooling.

[0070] BT i,1,k For workpiece J i The first process is on machine tool M k The start time of the upper processing.

[0071] Y i,j,k,r As a decision variable, if the process Q after batch processing... ij In machine tool M k Using tooling MT r If the value is 1, then the value is 0; otherwise, the value is 0.

[0072] R r As a decision variable, if tooling MT r If available at the initial scheduling time, the value is 1; otherwise, it is 0.

[0073] Furthermore, the specific implementation steps of the model solving method based on the knowledge-driven batch method and the improved multi-objective evolutionary algorithm in step S2 are as follows:

[0074] A two-stage feedback optimization approach is adopted, utilizing a knowledge-driven batching method framework to optimize the workpiece batching sub-problem. Based on the proposed improved multi-objective evolutionary algorithm, the sub-problems of machine tool allocation, process sequencing, and tooling allocation are further optimized. That is:

[0075] Workpiece batches are divided based on the equal batching method and the variable minimum batch constraint.

[0076] The performance of the proposed improved multi-objective evolutionary algorithm is evaluated on the workpiece batch partitioning scheme.

[0077] Keeping the initial batch partitioning scheme unchanged in each iteration, the local solution space of the problem is searched based on the number of workpiece batches.

[0078] Based on the different problem sizes under different batching schemes, the key parameters of the algorithm are dynamically and adaptively adjusted.

[0079] A three-vector encoding method and a greedy active decoding method are adopted to achieve a two-way mapping between the solution space and the problem space of the subproblems of machine tool allocation, process sequencing, and tooling allocation.

[0080] A method for mutual conversion between individual positions and each encoding vector is introduced to continuously process the energy-saving batch flow scheduling problem of flexible workshops that considers processing preparation resources, providing a basis for its subsequent optimization.

[0081] Based on the number of processes q for each workpiece i Generate an initial workpiece information sequence, and simultaneously generate sequences from 1 to 1 based on the total number of processes. The continuous numbers form the process number vector. Continuous values ​​for individual positions are randomly generated within the range [-α, α], where α > 0. Next, the generated continuous values ​​are arranged in descending order according to their size, while the process number vector corresponding to the position information is adjusted. Finally, the actual process sorting information is formed through process replacement.

[0082] The formula for converting location information into machine tool and tooling allocation information is as follows:

[0083] Where, m u Assign a specific process Q to a machine tool or tooling. ij Position vector, s u For use in process Q ij The number of machine tools or fixtures used for processing.

[0084] The inverse function of the above formula converts the continuous values ​​of the position vector into discrete values ​​of the machine tool and tooling allocation scheme.

[0085] The continuous values ​​of the position vector are converted into discrete values ​​of the process sequencing scheme using the following formula:

[0086]

[0087] in, r is the actual process sequencing information vector. u ∈r,

[0088] Under the knowledge-driven batching method framework, the following formula is used to initialize individual moths:

[0089] x u =LB+rand(0,1)×(UB-LB)

[0090] Where, x u For each individual moth, UB and LB are the upper and lower bounds of the individual component values, respectively, which are 1 and -1 here.

[0091] Based on the moth flame operator, the three-dimensional solution space of the energy-saving batch flow scheduling problem of flexible workshop considering processing preparation resources is fully explored.

[0092] The formula for moth search is: x u =dis(x) u ,f u )×e bt ×cos(2πt)+f u

[0093] Where, dis(x) u ,f u ) represents the moth x u With flame f u The Euclidean distance between them.

[0094] b is the defined helical shape constant, which is 0.01.

[0095] t is the path coefficient, and its value ranges from [-1, 1].

[0096] Multiple co-evolutionary operators based on genetic and differential evolution operators promote information exchange among individual moths.

[0097] Genetic operators mainly consist of two parts: crossover and mutation. The crossover operation is represented as:

[0098] w u =0.5×[(1+κ)×x u +(1-κ)×x v ]

[0099] w v =0.5×[(1+κ)×x u +(1-κ)×x v ]

[0100] The mutation operation is represented as:

[0101]

[0102] Differential evolution operators include differential mutation and crossover. The crossover operation is consistent with the crossover operation in genetic operators. There are three mutation operators, denoted as follows:

[0103] DE / rand / 1:W u =x a +F×(x b -x c )

[0104] DE / rand / 2:W u =x a +F×(x b -x c)+F×(x d -x e )

[0105] DE / current-to-rand / 1:

[0106] Among them, W u , and These are different offspring of the moth.

[0107] x v ,x a x b x c x d and x e These are different parent individuals of the moth.

[0108] K and F are scaling factors, which are set to 0.5 here.

[0109] η is the distribution factor.

[0110] △ max The maximum difference in the individual component values ​​of a moth.

[0111] A selection strategy based on the IGD-NS (i.e., inverse generation distance considering non-contributing solutions) index is introduced to obtain a multi-objective non-dominated solution set for the energy-saving batch flow scheduling problem of flexible workshop considering processing preparation resources, ensuring the continuous optimization capability of the moth population; the calculation formula of the IGD-NS index is:

[0112]

[0113] Where P' is the set containing all non-contributing solutions, P is the set of non-dominated solutions, and P * For reference set.

[0114] dis(x,y) represents the point y in the solution set P and the reference set P. * The Euclidean distance between points x in the matrix.

[0115] dis(x,y') represents the point y' in the solution set P' and the reference set P'. * The Euclidean distance between points x in the matrix.

[0116] Among them, a non-contribution solution refers to a non-dominated solution that is not the nearest neighbor of any point in the reference set.

[0117] Furthermore, the specific implementation steps of the tooling adjustment strategy based on adjacent processes in step S3 are as follows:

[0118] Step 1: Obtain the current status information of the workpiece, machine tool, and tooling.

[0119] Step 2: Determine the first process to begin among the selected processes. If multiple processes start simultaneously, randomly select one of them.

[0120] Step 3: Determine the adjacent processes of this process and the preceding process on the selected machine tool. Do you want to select the same tooling for processing? If yes, then discard this process and return to step 2; otherwise, proceed to the next step.

[0121] Step 4: Determine the process Can the tooling be used in the process? Its process Is it available within the processing cycle? If yes, then... Tooling and If the settings are consistent, update the tooling status information and proceed to the next step; otherwise, discard the process and return to step 2.

[0122] Step 5. End the algorithm.

[0123] in, For machine tool M k The j-th process in the upper processing, For machine tool M k The (j-1)th process in the upper processing.

[0124] Furthermore, step S4 includes the following sub-steps:

[0125] To verify the effectiveness of the energy-saving batch flow scheduling optimization method for flexible job shops proposed in this invention, the benchmark case 20×MK06 from the literature "Tian ZQ, Jiang XY, Liu WJ, et al. Knowledge-based lot-splitting optimization method for flexible job shops considering energy consumption[J]. IEEE Transactions on Automation Science and Engineering, early access, 2023" was selected for testing. This example includes 10 workpieces, 10 machine tools, an average of 15 processes per workpiece, an average of 3 machine tools that can be selected for processing per process, and an average batch size of 20 workpieces. The proposed knowledge-driven batching method and the improved multi-objective evolutionary algorithm hybrid method, MOEA / D, NSGA-II, MOPSO, MOEA / PSL, and LMOCSO were used to solve the problem. The statistical results of 10 solutions are shown in Table 1. As can be seen from Table 1, for the average value, the proposed knowledge-driven batching method and the improved multi-objective evolutionary algorithm hybrid method are significantly better than their comparative algorithms. The Gantt chart of the optimized scheduling scheme is shown below. Figure 3 and 4 As shown, the batch division scheme for each workpiece is as follows: [12, 10; 12, 17; 12, 8; 12, 9; 14, 19, 17, 10, 12, 12, 14]. The completion time of the workpiece, the total energy consumption of the workshop, and the total processing cost are 4092.9h, 362053kW·h, and 104635CNY, respectively.

[0126] Table 1. Statistical results of the benchmark case 20×MK06 under different solution algorithms.

[0127]

[0128] Furthermore, the example verification in step S5 comes from a production workshop of a machine tool manufacturing company. This workshop has 16 processing equipment, including gantry machining centers, horizontal CNC machine tools, medium-sized CNC machine tools, small CNC machine tools, vertical machining centers, and boring and milling machining centers. A batch of orders processed bed, column, and slide parts in batches of 10, 15, and 20 respectively, with 22, 15, and 18 operation steps for each part, and 9 available tooling fixtures. The proposed knowledge-driven batch processing method and the improved multi-objective evolutionary algorithm hybrid method, MOEA / D, NSGA-II, MOPSO, MOEA / PSL, and LMOCSO were used to solve the problem. The comparison results of each index after 10 solutions are as follows: Figure 5(a) , 5(b) As shown in 5(c). From Figure 5(a) , 5(b) As shown in Figure 5(c), the proposed hybrid method combining knowledge-driven batch processing and improved multi-objective evolutionary algorithm significantly outperforms its comparative algorithms in all metrics. The Gantt charts of the optimized proposed method and the enterprise's current scheduling scheme are as follows: Figure 6 and 7 As shown, the current scheduling scheme for the enterprise results in a workpiece completion time of 2778.23 min, a total workshop energy consumption of 4035.66 kW·min, and a total processing cost of 60916.9 CNY. The optimized scheduling scheme based on the proposed method results in a workpiece completion time of 1872.3 min, a total workshop energy consumption of 3731.73 kW·min, and a total processing cost of 58666.1 CNY. It can be seen that after optimization using the proposed method, the workpiece completion time is reduced by 32.6%, the total workshop energy consumption is reduced by 7.5%, and the total processing cost is reduced by 3.7%.

[0129] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it. They should not be used to limit the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the essence of the present invention shall fall within the scope of protection of the present invention.

Claims

1. A multi-objective optimization method for energy-saving batch flow scheduling in flexible workshops, considering processing preparation resources, characterized in that, Includes the following steps: S1. Multi-objective optimization mathematical model for energy-saving batch flow scheduling in flexible workshops considering processing preparation resources; S2. A model solution method based on knowledge-driven batch processing and an improved multi-objective evolutionary algorithm; S3. Tooling adjustment strategy based on adjacent processes. The specific implementation steps of the model solving method based on the knowledge-driven batch method and the improved multi-objective evolution algorithm in step S2 are as follows: A two-stage feedback optimization approach is adopted, utilizing a knowledge-driven batching method framework to optimize the workpiece batching sub-problem. Based on the proposed improved multi-objective evolutionary algorithm, the sub-problems of machine tool allocation, process sequencing, and tooling allocation are further optimized. Workpiece batches are divided based on the equal batching method and the variable minimum batch constraint. The performance of the proposed improved multi-objective evolutionary algorithm is evaluated to assess the workpiece batch partitioning scheme. Keeping the initial batch partitioning scheme unchanged in each iteration, the local solution space of the problem is searched based on the number of workpiece batches. Based on the different problem sizes under different batching schemes, the key parameters of the algorithm are dynamically and adaptively adjusted. A three-vector encoding method and a greedy active decoding method are adopted to achieve a bidirectional mapping between the solution space and the problem space of the subproblems of machine tool allocation, process sequencing, and tooling allocation. A method for converting between individual positions and encoded vectors is introduced to continuously process the energy-saving batch flow scheduling problem of flexible job shops that considers processing preparation resources, providing a foundation for its subsequent optimization. Based on the number of processes q for each workpiece i Generate an initial workpiece information sequence, and simultaneously generate sequences from 1 to 1 based on the total number of processes. The process number vector is formed by continuously generating consecutive values ​​for individual positions within the range [-α, α], where α >

0. Next, the generated consecutive values ​​are sorted in descending order based on their magnitude, while simultaneously adjusting the process number vector corresponding to the position information. Finally, the actual process sequence information is formed through process permutation. The formula for converting location information into machine tool and tooling allocation information is as follows: Where, m u Assign a specific process Q to a machine tool or tooling. ij Position vector, s u For use in process Q ij The number of machine tools or fixtures used for processing. The inverse function of the above formula converts the continuous values ​​of the position vector into discrete values ​​of the machine tool and tooling allocation scheme. The continuous values ​​of the position vector are converted into discrete values ​​of the process sequencing scheme using the following formula: in, This is a vector of actual process sequencing information. Under the knowledge-driven batching method framework, the following formula is used to initialize individual moths: x u =LB+rand(0,1)×(UB-LB) Where, x u For each individual moth, UB and LB are the upper and lower bounds of the individual component values, respectively, which are set to 1 and -1 here. Based on the moth flame operator, the three-dimensional solution space of the energy-saving batch flow scheduling problem in flexible workshops, considering processing preparation resources, is fully explored. The formula for moth search is: x u =dis(x) u ,f u )×e bt ×cos(2πt)+f u Where, dis(x) u ,f u ) represents the moth x u With flame f u Euclidean distance between them b is the defined spiral shape constant, taken as 0.01, and t is the path coefficient, with a value range of [-1, 1]. Multiple co-evolutionary operators based on genetic and differential evolution operators promote information exchange among individual moths. Genetic operators mainly consist of two parts: crossover and mutation. The crossover operation is represented as: w u =0.5×[(1+κ)×x u +(1-k)×x v ] w v =0.5×[(1+κ)×x u +(1-k)×x v ] The mutation operation is represented as: Differential evolution operators include differential mutation and crossover. The crossover operation is consistent with the crossover operation in genetic operators. There are three mutation operators, denoted as follows: DE / row / 1:W u =x a +F×(x b -x c ) DE / rand / 2:W u =x a +F×(x b -x c )+F×(x d -x e ) DE / current-to-rand / 1: Among them, W u , and These are different offspring of the moth. x v ,x a x b x c x d and x e These are different parent individuals of the moth. K and F are scaling factors, which are set to 0.5 here. η is the distribution factor. Δ max The maximum difference in the individual component values ​​of the moth. A selection strategy based on the IGD-NS index is introduced to obtain a multi-objective non-dominated solution set for the energy-saving batch flow scheduling problem of flexible workshops, considering processing preparation resources, thus ensuring the continuous optimization capability of the moth population. The calculation formula for the IGD-NS index is as follows: Where P′ is the set containing all non-contributing solutions, P is the set of non-dominated solutions, and P * For reference set, dis(x,y) represents the point y in the solution set P and the reference set P. * The Euclidean distance between points x in the solution set P′ and the reference set P′ is represented by dis(x,y′). * The Euclidean distance between points x in the equation. Here, a non-contribution solution refers to a non-dominated solution that is not the nearest neighbor of any point in the reference set. The specific implementation steps of the tooling adjustment strategy based on adjacent processes in step S3 are as follows: Step 1: Obtain the current status information of the workpiece, machine tool, and fixture. Step 2: Determine the first process to begin among the selected processes. If multiple processes start simultaneously, randomly select one of them. Step 3: Determine the adjacent processes of this process and the preceding process on the selected machine tool. If you choose to process the same type of tooling, then cancel the process and return to step 2; otherwise, proceed to the next step. Step 4: Determine the process Can the tooling be used in the process? Its process Is it available within the processing cycle? If yes, then... Tooling and If the settings are consistent, update the tooling status information and proceed to the next step; otherwise, discard the process and return to step 2. Step 5. End the algorithm. in, For machine tool M k The j-th process in the upper processing, For machine tool M k The (j-1)th process in the upper processing.

2. The multi-objective optimization method for energy-saving batch flow scheduling in flexible workshops considering processing preparation resources as described in claim 1, characterized in that, A multi-objective optimization mathematical model for energy-saving batch flow scheduling in flexible job shops, considering processing preparation resources, includes: The optimization objectives of the model are to minimize the completion time, total energy consumption, and total processing cost. In addition to meeting the basic constraints of batch flow scheduling in flexible workshops, it is also necessary to satisfy the constraint that the same tooling can only be applied to one process operation at a time, as expressed as: Among them, ST i,j,r For process Q ij Occupying tooling MT r Start time; RT g,h,r For process Q ij Occupying tooling MT r End time; As a decision variable, if process Q ij With subsequent processes Q gh All require the use of tooling MT r The value is 1 if the condition is met, otherwise it is 0; B is a sufficiently large real number; O i,j For the machinable process Q ij Tooling MT r Set; O g,h For the machinable process Q gh Tooling MT r Set; n′ is the total number of workpieces after batch partitioning; i, g are workpiece indices; h is the process index; r is the tooling index; q i For workpiece J i Number of processes; q g For workpiece J g The number of processes, In addition, the constraint that all tooling is available at the initial moment of scheduling must also be satisfied, expressed as: Y i,j,k,r ∈{0,1},i∈[1,n′],j∈[1,q i ],k∈[1,m],r∈[1,N t ] R r ∈{0,1}, Where k is the machine tool index; m is the number of machine tools; N t For tooling quantity; BT i,1,k For workpiece J i The first process is on machine tool M k Start time of the upper processing; Y i,j,k,r As a decision variable, if the process Q after batch processing... ij In machine tool M k Using tooling MT r R takes 1 if the condition is met, otherwise takes 0; r As a decision variable, if tooling MT r If available at the initial scheduling time, the value is 1; otherwise, it is 0.

Images