Cross-unit fuzzy scheduling method based on ant lion optimization algorithm

Through the cross-unit fuzzy scheduling method based on the Ant Lion optimization algorithm, triangular fuzzy numbers and three-segment coding, the problem of production time uncertainty in the existing technology is solved. The generated scheduling scheme is more accurate and adaptable, and effectively guides the actual situation. Production.

CN118917620APending Publication Date: 2024-11-08CHONGQING UNIV
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411114612.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing cross-unit production scheduling method assumes that the production time is accurate and cannot effectively reflect the actual processing conditions, resulting in production delays and product delivery delays, and loses its guiding role in actual production.

Method used

Using a cross-unit fuzzy scheduling method based on the ant lion optimization algorithm, the workpiece processing and transportation time is characterized by triangular fuzzy numbers, and a scheduling model that minimizes the maximum fuzzy completion time is built, and an effective scheduling solution is generated using three-stage encoding and ant lion optimization algorithm. .

Benefits of technology

The accuracy and robustness of the scheduling model are improved, the generated scheduling scheme is more in line with the actual production situation, enhances the adaptability and robustness of the system, and can effectively guide actual production.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917620A_ABST
    Figure CN118917620A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of cross-unit production, and particularly relates to a cross-unit fuzzy scheduling method based on an ant lion optimization algorithm, and the method comprises the steps: S1, representing the fuzzy processing time and the fuzzy transportation time in a workpiece processing process through a triangular fuzzy number; by taking minimization of fuzzy maximum completion time as an optimization target, constructing a scheduling model of cross-unit scheduling production; s2, processing data information and cross-unit transportation data information of the workpiece are acquired, and all possible potential scheduling schemes are generated; s3, performing three-section type coding on each potential scheduling scheme based on process sorting, machine selection and manufacturing unit selection; s4, solving the scheduling model by using an ant lion optimization algorithm to obtain an optimal scheduling scheme; and S5, performing actual cross-unit scheduling production by using the optimal scheduling scheme obtained in the step S4. According to the method, an effective cross-unit scheduling scheme can be obtained, it is guaranteed that the scheme conforms to the actual machining condition, and actual production can be effectively guided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of cross-unit production, and in particular relates to a cross-unit fuzzy scheduling method based on an ant lion optimization algorithm. Background Art

[0002] Due to the increasing degree of customization in current market demand, the discrete manufacturing industry mostly adopts a multi-variety, small-batch production organization model, and its production and manufacturing processes require higher flexibility and adaptability. Cellular Manufacturing, as a modern production method that integrates group technology into production, is widely used in the production of discrete manufacturing enterprises. That is, the manufacturing workshop is divided into independent units of different sizes, each unit is responsible for completing a specific production task or production link, which reduces the workpiece transfer time, reduces the level of work-in-process inventory, and reduces the workpiece production preparation time to a certain extent.

[0003] However, on the one hand, manufacturing workshops are facing the production requirements of more and more diversified and segmented products and are carrying out customized production; on the other hand, some special manufacturing workshops, such as R&D workshops, need to take into account both trial production and mass production. Since customized production and trial production usually require more production flexibility and adjustment capabilities, while mass production focuses more on high efficiency and stability, the three use the same production resources in the manufacturing workshop, which results in a large number of workpieces or products in the current production tasks that are significantly different from the part family of the configured manufacturing unit (i.e., abnormal parts). In order to effectively avoid the loss of production costs and equipment processing accuracy caused by the reconstruction of manufacturing unit configuration and the purchase of new equipment, the cross-unit production strategy has become a necessary choice for unit production of discrete manufacturing enterprises today. It realizes collaborative production and resource sharing of manufacturing units by reasonably allocating equipment resources.

[0004] In order to realize cross-unit production more reasonably and effectively, more advanced production scheduling technology is needed as support. The key basis for production scheduling is the processing time and transportation time of the workpiece. Among them, the processing time in the production scheduling problem is usually composed of production preparation time (such as machine tool debugging time, workpiece clamping time), workpiece cutting time, quality inspection time and part loading and unloading time. Although with the rapid development of CNC technology, the cutting work of the workpiece can be completed on time, but work such as machine tool debugging, clamping, disassembly, and measurement are still done manually. The workers' operating proficiency and operating efficiency greatly affect the accuracy of this type of time, making it impossible to maintain the processing time of the workpiece at an exact value. In addition, the transportation time is also composed of multiple time periods such as loading, transportation, and disassembly of the workpiece. Among them, the loading and disassembly work is still done manually, and there is traffic congestion in the transportation process, and the transportation time cannot be accurately estimated.

[0005] The above two types of time can both be called production time in cross-unit production. At present, a large number of studies on production scheduling assume that the production time is an accurate value. This assumption is seriously inconsistent with the actual processing situation, which seriously affects the execution rate of the production scheduling plan, resulting in production delays and product delivery delays, causing the scheduling plan to lose its guiding role in actual production.

[0006] Therefore, how to obtain an effective cross-unit scheduling plan to ensure that it is consistent with the actual processing situation and can provide effective guidance for actual production has become a problem that needs to be solved urgently. Summary of the invention

[0007] In view of the above-mentioned deficiencies in the prior art, the present invention provides a cross-unit fuzzy scheduling method based on the ant lion optimization algorithm. The method can obtain an effective cross-unit scheduling scheme, ensure that it is consistent with the actual processing situation, and can effectively guide actual production.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0009] A cross-unit fuzzy scheduling method based on the ant lion optimization algorithm comprises the following steps:

[0010] S1. Use triangular fuzzy numbers to characterize the fuzzy processing time and fuzzy transportation time in the workpiece processing process; take minimizing the fuzzy maximum completion time as the optimization goal, and build a scheduling model for cross-unit scheduling production;

[0011] S2, obtaining the processing data information and cross-unit transportation data information of the workpiece, and generating all possible potential scheduling schemes; the processing data information includes parts, processes, optional equipment and corresponding fuzzy processing time; the cross-unit transportation data information includes fuzzy transportation time;

[0012] S3, three-stage coding of each potential scheduling scheme based on process sequencing, machine selection, and manufacturing unit selection;

[0013] S4. Based on the three-stage coding of each potential scheduling solution, the scheduling model is solved using the antlion optimization algorithm to obtain the best scheduling solution;

[0014] S5. Use the best scheduling solution obtained in S4 to carry out actual cross-unit production scheduling.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] 1. This method uses triangular fuzzy numbers to characterize the fuzzy processing time and fuzzy transportation time in the workpiece processing process, which can more accurately reflect the uncertainty in actual production. Triangular fuzzy numbers can fully describe the fuzziness and uncertainty range of time variables by defining three parameters: minimum value, most likely value, and maximum value. This representation method enables the scheduling model to more realistically reflect the actual situation in the production process, thereby improving the accuracy and reliability of the model.

[0017] The scheduling model takes minimizing the fuzzy maximum completion time as the optimization goal. By introducing fuzzy number representation and fuzzy optimization objectives, the scheduling model can better cope with uncertain factors in the production process (such as equipment failure, raw material delays, etc.) and improve the robustness and adaptability of production scheduling. Since the scheduling model takes into account the uncertainty and multi-objective optimization requirements in the production process, the generated scheduling plan is more in line with the actual production situation, which helps enterprises improve decision-making efficiency and production benefits. The use of fuzzy time representation and ant lion optimization algorithm makes the scheduling plan have a certain adaptability to changes in the external environment, enhancing the robustness of the system.

[0018] 2. Considering the discrete optimization characteristics of the cross-unit scheduling problem with fuzzy time, this method encodes each potential scheduling scheme in three stages based on process sorting, machine selection, and manufacturing unit selection, respectively, and then solves it through the ant lion optimization algorithm after corresponding discretization improvements, which can ensure the effectiveness of the obtained scheduling scheme.

[0019] 3. The three-stage coding decomposes the scheduling problem into three sub-problems: process sorting, machine selection, and manufacturing unit selection. By encoding each sub-problem separately, the key information in the scheduling plan can be more clearly represented, which is convenient for subsequent optimization algorithms to process. In addition, in the Ant Lion optimization algorithm, the use of three-stage coding can narrow the search space, allowing the algorithm to find high-quality solutions more quickly. Since the encoding of each sub-problem is relatively independent, the algorithm can perform a more detailed search in the solution space of each sub-problem, thereby improving search efficiency.

[0020] In addition, the three-stage coding ensures that the generated scheduling plan meets the constraints in actual production (such as dependencies between processes, capacity limitations of equipment, etc.) through clear coding rules. This coding method enables the algorithm to automatically exclude invalid solutions that do not meet the constraints during the search process, thereby improving the feasibility of the final solution. Moreover, the simplicity and clarity of the three-stage coding make the implementation of the algorithm easier. Developers can design corresponding algorithm steps and data structures according to the coding rules to achieve efficient scheduling optimization. At the same time, the three-stage coding is also convenient for debugging and verification of the algorithm, because the coding of each sub-problem is relatively independent, so the correctness of the algorithm can be verified by testing each sub-problem separately.

[0021] In summary, this method can obtain an effective cross-unit scheduling plan, ensure that it is consistent with the actual processing situation, and can provide effective guidance for actual production.

[0022] Preferably, in S4, the process of solving the scheduling model by the ant lion optimization algorithm includes:

[0023] Step 1: Initialize algorithm parameters, including the ant population size P a , the antlion population size P L , the maximum number of iterations of the algorithm maxGen;

[0024] Step 2: Generate P using random initialization method a +P L Three-part coding of each individual;

[0025] Step 3: Greedily decode the three-segment code of each individual to obtain the maximum fuzzy completion time and defuzzify it. Then, the reciprocal of the defuzzified value is taken as the fitness value of the individual.

[0026] Step 4: Sort the individuals according to their fitness values ​​from large to small, and sort the top P L The number of individuals is regarded as the ant lion population, and the remaining P a The number of individuals is regarded as the ant population; and the individual with the largest fitness value is regarded as the elite ant lion;

[0027] Step 5: Each ant performs random walks and updates its position according to the current state of the antlion population to achieve global optimization;

[0028] Step 6: Construct the neighborhood solution of each antlion according to the preset method and perform local optimization;

[0029] Step 7: Combine the results of Step 5 and Step 6 to update the position of each ant lion and re-determine the elite ant lion;

[0030] Step 8: According to the preset method, the re-determined elite ant lion is subjected to variable neighborhood search to obtain the final elite ant lion of the current iteration;

[0031] Step 9: Determine whether the number of iterations reaches the preset maximum number of iterations maxGen. If not, return to step 5; if reached, take the scheduling plan corresponding to the final elite ant lion of the current iteration as the best scheduling plan.

[0032] Such a setting,1,uses three-segment coding to represent the scheduling scheme,and obtains the maximum fuzzy completion time through greedy decoding,and then converts it into fitness value using defuzzification technology.,This process is simple and efficient.,The calculation method of fitness value (i.e., the inverse of the maximum fuzzy completion time) can,intuitively reflect the quality of the solution, which is convenient for,algorithm selection and comparison.

[0033] 2. The algorithm can handle complex cross-unit scheduling problems involving fuzzy time and three-segment coding. By introducing fuzzy number representation and three-segment coding technology, the algorithm can more accurately describe the uncertainty factors in actual production and improve the adaptability and practicality of the scheduling model.

[0034] 3. The Ant Lion Optimization Algorithm implements global search through the random walk and position update of ants, which can explore a wide range of areas in the solution space and help discover new potential solutions. At the same time, the algorithm can perform a detailed search near the discovered high-quality solutions by constructing the neighborhood solution of the Ant Lion and performing local optimization, thereby improving the accuracy of the solution. This combination of global and local search enables the algorithm to quickly converge to a high-quality solution while maintaining search diversity.

[0035] 4. The elite strategy ensures that the algorithm continuously evolves to a better solution during the iteration process by retaining and optimizing the individuals with the highest fitness values ​​(elite ant lions). The variable neighborhood search further optimizes the elite ant lions, avoiding falling into local optimality by exploring different neighborhood structures, and improving the global search capability of the algorithm.

[0036] Preferably, the expression of the defuzzification operation is:

[0037]

[0038] In the formula, Represents triangular fuzzy time The defuzzified value of a1 represents the earliest process completion time or the workpiece cross-unit transportation time, M It represents the most likely process completion time or the workpiece cross-unit transportation time, and a2 represents the latest process completion time or the workpiece cross-unit transportation time.

[0039] Preferably, in step 5, the process of ants performing full optimization includes:

[0040] Select an antlion for each ant via roulette or other selection mechanism;

[0041] According to the dynamic coefficient a, the ant individual is selected to cross with the ant lion selected by the roulette wheel, or cross with the elite ant lion to update the position of the ant individual; the formula for updating the position of the ant individual is:

[0042]

[0043] In the formula, Ant i (t+1) indicates the position of the ant in the t+1th iteration; i (t) represents the position of the antlion selected by ant individual i in the roulette wheel in the tth iteration; BestAntlion(t) represents the position of the elite antlion in the tth iteration; is a crossover symbol, indicating that the codes of individual A and individual B are cross-operated; rand is a random value in the interval [0,1]; a is a dynamic coefficient that changes with the current number of iterations, and t is the current number of iterations;

[0044] The dynamic coefficient calculation formula is:

[0045]

[0046] Where t is the current iteration number and T is the maximum iteration number.

[0047] Such a setting, 1. By modifying the random walk rule of ants to adapt it to the search of discrete variables, ants can effectively search in the discrete solution space. The search behavior of ants is made more consistent with the actual constraints of the scheduling problem, improving the feasibility of the solution. By modifying the random walk rule of ants to a roulette-based selection of ant lions and crossover, the algorithm is more suitable for the search problem of discrete variables. This modification enables the algorithm to explore more effectively in the discrete solution space when solving the scheduling problem, improving the adaptability of the algorithm to practical problems.

[0048] 2. It can improve the diversity of solutions. The roulette selection mechanism provides each ant with the opportunity to choose a different antlion, increasing the diversity of individuals in the ant population. At the same time, the introduction of the dynamic coefficient changes the strategy of ants in selecting crossover objects at different iteration stages, further promoting the diversity of solutions in the population. This diversity helps the algorithm to jump out of the local optimal solution and explore a wider solution space.

[0049] 3. The setting of the dynamic coefficient α can adjust the selection strategy of ant individuals in the global search process to prevent elite ant lions from falling into the local optimal solution too early in the early iteration. By dynamically adjusting the value of α, the algorithm can balance the search breadth and depth of ants, ensuring that ants can effectively explore a better solution space while maintaining a certain degree of randomness.

[0050] Preferably, when the ant individual is crossed with the roulette wheel selected ant lion or the elite ant lion, the machine selection code and the unit selection code use joint uniform crossover, and the process sorting code uses IPOX crossover.

[0051] Such a setting can improve the adaptability of the encoding type. Machine selection encoding and unit selection encoding usually belong to integer encoding or combination encoding, which often involves a limited set of discrete choices. The use of joint uniform crossover can ensure that during the crossover process, the corresponding coding parts of the two parent individuals can exchange genes with equal probability, thereby maintaining the diversity of solutions and potentially producing better offspring. Process sorting encoding involves the order of processes, which is crucial for scheduling problems. IPOX (Iterated Pairwise Order Crossover) crossover is a crossover operator specifically used for sequence problems. It can effectively maintain the relative order relationship between processes, reduce illegal solutions generated during the crossover process, and thus improve the quality of the solution.

[0052] It can also enhance the global search capability of the algorithm. Ant lions or elite ant lions selected by roulette are usually individuals with better performance in the current population. Crossing with these individuals helps ant individuals absorb the characteristics of excellent individuals, thereby accelerating the evolution of the entire population. Different types of encoding use different crossover strategies, which can more accurately handle the characteristics of various encoding types and avoid gene damage or solution quality degradation caused by improper crossover methods. This helps to improve the global search capability of the algorithm, enabling it to search for high-quality solutions in a wider solution space.

[0053] In addition, local search and neighborhood exploration can be promoted. After the crossover operation, a series of local search or neighborhood exploration steps are usually followed to further optimize the quality of the solution. The offspring generated by the hybrid crossover strategy not only inherit the excellent characteristics of the parent individuals, but also introduce new gene combinations through crossover, which provides more possibilities and directions for subsequent local searches. Especially for the case where the process sorting encoding adopts IPOX crossover, since IPOX crossover can maintain the relative order relationship between processes, the generated offspring are more likely to maintain the legitimacy of the solution during the local search process, thereby reducing unnecessary repair operations and computational costs.

[0054] At the same time, since different types of encodings use appropriate crossover methods, the algorithm is more stable during the search process and is not prone to falling into local optimal solutions.

[0055] Preferably, in step 6, the process of constructing the neighborhood solution of each ant lion includes:

[0056] For the individual ant lion machine, select the MS code and the unit select the CS code; randomly select 2-3 coding positions, determine the machine number with the minimum processing time in the processing equipment concentration of the coding position process, and replace the original coding position; the CS code is corrected and adjusted as the MS code changes;

[0057] For the process selection OS code of the antlion individual, three process code positions are randomly selected and exchanged with each other to generate 6 sections of neighborhood process selection codes;

[0058] The 6-segment neighborhood process selection codes are combined with the adjusted MS code and CS code to obtain 6 groups of neighborhood solutions;

[0059] Compare the fitness values ​​of the antlion individual and its six groups of neighborhood solutions, and take the position with the largest fitness value as the local optimization result of the antlion.

[0060] This setting can enhance local search capabilities. By randomly selecting 2-3 positions in the machine selection MS code and the unit selection CS code, replacing them based on the principle of optimizing the process processing time, and exchanging the three process positions in the process selection OS code, multiple sets of neighborhood solutions are generated. This method effectively conducts a detailed search in the neighborhood of the ant lion individual. This local search can help the algorithm find better solutions near the discovered high-quality solutions, thereby accelerating the convergence process.

[0061] 2. It can improve the diversity of solutions. Although the search is conducted in a local range, the random selection of encoding positions and process exchange methods still maintains the diversity of the search. This diversity helps the algorithm avoid falling into the local optimal solution and increases the possibility of finding the global optimal solution.

[0062] 3. When the MS code changes, the CS code can be automatically corrected and adjusted to ensure the legitimacy of the solution. This linkage mechanism ensures the integrity and validity of the solution during the search process and avoids illegal or invalid solutions caused by inconsistent codes.

[0063] 4. By constructing the neighborhood solution of the ant lion and performing local optimization, the overall performance of the algorithm has been significantly improved. The strategy of combining local search with global search enables the algorithm to perform fine search in the local range while maintaining the global search capability, thereby improving the solution quality and efficiency of the algorithm.

[0064] Preferably, in step 8, the process of changing the neighborhood search includes:

[0065] Step 8.1: Take the elite ant lion as the initial solution X and set the maximum number of neighborhood searches q max , and set the parameters q=0, k=1;

[0066] Step 8.2: Use the neighborhood structure N for the initial solution X k (k=1,2…,n) performs local search. If the fitness f(X') of the generated new individual X' is greater than the fitness f(X) of the original individual, the original individual is replaced by X' and k=1; otherwise, k=k+1; where Nk represents the kth preset neighborhood structure; n is the type of the preset neighborhood structure;

[0067] Step 8.3: Determine whether k>n. If so, go to step 8.4; otherwise, go to step 8.2;

[0068] Step 8.4: Let q = q + 1, and determine whether the neighborhood search termination condition q>q is satisfied max If it is satisfied, jump to step 8.5; if it is not satisfied, set k=1 and go to step 8.2;

[0069] Step 8.5: The algorithm ends and the result of the neighborhood search is output as the final elite antlion of the current iteration.

[0070] Such a setting, 1. Enhance the global search capability. Although variable neighborhood search is a local search technology, by constantly switching the neighborhood structure (N k ) and restart the search from the elite antlion. The algorithm can jump out of the local optimal solution to a certain extent and increase the breadth of the global search. This helps to avoid premature convergence of the algorithm and increase the possibility of finding the global optimal solution.

[0071] 2. Improve system flexibility: By introducing a variety of preset neighborhood structures (N k ), the algorithm can flexibly adjust the search strategy according to the characteristics of the problem and the current search status, enhancing the adaptability and flexibility of the system. This is of great significance for dealing with complex and changing manufacturing environments.

[0072] 3. Improve the accuracy of the solution. By continuously performing local searches in the neighborhood of the elite antlion (i.e., the current optimal solution), the algorithm can try multiple possible solution space regions, thus having the opportunity to find a solution that is better than the current optimal solution. This sophisticated search strategy helps improve the accuracy and performance of the solution.

[0073] 4. Accelerate the convergence speed. Although variable neighborhood search increases the computational complexity of the algorithm, it also accelerates the convergence speed of the algorithm to a certain extent. Because during the search process, the algorithm can more quickly identify and focus on the potential optimal solution area, reducing unnecessary search space exploration.

[0074] Preferably, the triangular fuzzy number representing the fuzzy processing / transportation time is In the formula, represents the fuzzy processing / transportation time; x1 represents the earliest process / transportation completion time; x M represents the most likely process / transportation completion time; x2 represents the latest process / transportation completion time.

[0075] Thus, by explicitly specifying the earliest, most likely, and latest process / transportation completion times, triangular fuzzy numbers provide a more comprehensive description of the time range. This detailed representation enables the model to more accurately reflect the actual situation, thereby improving the accuracy and reliability of the model.

[0076] Preferably, the objective function of the scheduling model is:

[0077]

[0078] In the formula, Indicates workpiece J i The fuzzy completion time of the task; n is the number of workpieces.

[0079] Preferably, the constraints of the scheduling model include:

[0080]

[0081]

[0082] In the formula, Indicates workpiece J i The fuzzy completion time of the j-th process; Indicates workpiece J i The fuzzy start time of the j+1th process; N represents the set of workpiece serial numbers; O i For workpiece J i A set of processes;

[0083] Indicates workpiece J i The jth process is on machine M k The fuzzy completion time on Indicates workpiece J i The jth process is on machine M k Fuzzy start time on Y ijk When process O ij On device M k 1 when processing, otherwise 0; Indicates workpiece J i The jth process is on machine M k Fuzzy processing time on M ij For workpiece J i The set of processing equipment for the j-th process;

[0084] Indicates manufacturing unit M k With M k′ The fuzzy transportation time between ijk When process O ij On device M kIt is 1 when processing, otherwise it is 0; m represents the total number of processing equipment.

[0085] In this way, the first two formulas represent the processing order constraints of the previous and subsequent processes of the workpiece; the third formula indicates that each process can only start processing after the previous process is completed and transported to the manufacturing unit of the next process; the fourth formula indicates that the machine can only perform processing operations of one process at any time; the fifth formula indicates that all processes can only be processed on one machine at any time. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0087] Figure 1 A flowchart of the method is shown in FIG.

[0088] Figure 2 This is a schematic diagram of the three-segment encoding in Example 1;

[0089] Figure 3 This is the flow chart of the antlion optimization algorithm in this method;

[0090] Figure 4 This is a schematic diagram of the global search of individual ants in Embodiment 1 of the present method;

[0091] Figure 5 This is a schematic diagram of the local search of an ant lion individual in Example 1 of the present method;

[0092] Figure 6 This is a schematic diagram of the neighborhood structure in Embodiment 1 of the present method;

[0093] Figure 7 This is a processing information diagram of a production example part in Embodiment 2 of the present method;

[0094] Figure 8 This is the fuzzy scheduling Gantt chart in the second embodiment of the method. DETAILED DESCRIPTION

[0095] The following is a further detailed description through specific implementation methods:

[0096] Embodiment 1

[0097] like Figure 1 As shown, this embodiment discloses a cross-unit fuzzy scheduling method based on the ant lion optimization algorithm, including the following steps:

[0098] S1. Use triangular fuzzy numbers to represent the fuzzy processing time and fuzzy transportation time in the workpiece processing process; take minimizing the fuzzy maximum completion time as the optimization goal, and build a scheduling model for cross-unit scheduling production.

[0099] Among them, the triangular fuzzy number representing the fuzzy processing / transportation time is In the formula, represents the fuzzy processing / transportation time; x1 represents the earliest process / transportation completion time; x M represents the most likely process / transportation completion time; x2 represents the latest process / transportation completion time. By explicitly specifying the earliest, most likely, and latest process / transportation completion times, triangular fuzzy numbers provide a more comprehensive description of the time range. This detailed representation method enables the model to more accurately reflect the actual situation, thereby improving the accuracy and reliability of the model.

[0100] In specific implementation, the objective function of the scheduling model is:

[0101]

[0102] In the formula, Indicates workpiece J i The fuzzy completion time of the task; n is the number of workpieces.

[0103] The constraints of the scheduling model include:

[0104]

[0105] In the formula, Indicates workpiece J i The fuzzy completion time of the j-th process; Indicates workpiece J i The fuzzy start time of the j+1th process; N represents the set of workpiece serial numbers; O i For workpiece J i A set of processes;

[0106] Indicates workpiece J i The jth process is on machine M k The fuzzy completion time on Indicates workpiece J i The jth process is on machine M k Fuzzy start time on Y ijk When process O ij On device M k 1 when processing, otherwise 0; Indicates workpiece J i The jth process is on machine M k Fuzzy processing time on M ij For workpiece J i The set of processing equipment for the j-th process;

[0107] Indicates manufacturing unit Mk With M k, The fuzzy transportation time between ijk When process O ij On device M k It is 1 when processing, otherwise it is 0; m represents the total number of processing equipment.

[0108] Among the constraints, the first two formulas represent the processing order constraints of the previous and subsequent processes of the workpiece; the third formula indicates that each process can only start processing after the previous process is completed and transported to the manufacturing unit of the next process; the fourth formula indicates that the machine can only perform processing operations of one process at any time; the fifth formula indicates that all processes can only be processed on one machine at any time.

[0109] S2. Obtaining the processing data information and cross-unit transportation data information of the workpiece, and generating all possible potential scheduling schemes; the processing data information includes parts, processes, optional equipment and corresponding fuzzy processing time; the cross-unit transportation data information includes fuzzy transportation time.

[0110] S3. Encode each potential scheduling scheme in three stages based on process sequencing, machine selection, and manufacturing unit selection.

[0111] To describe this encoding method more clearly, a complete encoding of a simple example is given below. The example includes three workpieces to be processed. Each workpiece has three processes, and all workpieces are abnormal parts. The workpieces are respectively assigned to three manufacturing units for processing; the equipment serial numbers of the manufacturing units are divided as follows: equipment M1, M4 are located in manufacturing unit C1; M2, M6 are located in manufacturing unit C2; M3, M5 are located in manufacturing unit C3.

[0112] like Figure 2 As shown in the figure, the codes of the process sequence code part: the numbers 1, 2, and 3 represent the workpiece serial number, and the same numbers represent different processes of the same workpiece. The coding sequence increases from left to right. For example, the 3 on the process sequence code position 1 represents the first process of workpiece 3 (0 31 ), and the 3 on the second process code position represents the second process of workpiece 3 (O 32 ).

[0113] The machine selection code part and the unit selection code part are sequentially coded, that is, the codes are arranged from left to right according to the workpiece number and process number. For example, the codes in the first three positions of the machine code and the unit selection code represent the machine selection of process 1-3 of workpiece 1 and the unit number of the machine selected in the process, and the following fragments are analogous; 11There are two optional machines, M1 and M4. The corresponding machine selection code number 1 represents machine M1. The number 1 in the unit selection code is based on the fact that M1 is located in manufacturing unit C1. For process O that needs to be processed across units, 13 The optional machines include M2 ​​and M5. The corresponding machine code number 1 represents process O. 13 The second machine (M2) in the optional machine set is used for processing. Since M2 is located in the manufacturing unit C3, the number 3 is used in the unit selection code to indicate the unit where this process is processed.

[0114] S4. Based on the three-segment coding of each potential scheduling scheme, the scheduling model is solved using the antlion optimization algorithm to obtain the optimal scheduling scheme.

[0115] like Figure 3 As shown in the figure, the process of solving the scheduling model by the ant lion optimization algorithm includes:

[0116] Step 1: Initialize algorithm parameters, including the ant population size P a , the antlion population size P L , the maximum number of iterations of the algorithm maxGen;

[0117] Step 2: Generate P using random initialization method a +P L Three-part coding of each individual;

[0118] Step 3: Greedily decode the three-segment code of each individual to obtain the maximum fuzzy completion time and defuzzify it. Then, the reciprocal of the defuzzified value is taken as the fitness value of the individual.

[0119] In specific implementation, the expression of the defuzzification operation is:

[0120]

[0121] In the formula, Represents triangular fuzzy time The defuzzified value of a1 represents the earliest process completion time or the workpiece cross-unit transportation time, M It represents the most likely process completion time or the workpiece cross-unit transportation time, and a2 represents the latest process completion time or the workpiece cross-unit transportation time.

[0122] Step 4: Sort the individuals according to their fitness values ​​from large to small, and sort the top P L The number of individuals is regarded as the ant lion population, and the remaining P a The number of individuals is regarded as the ant population; and the individual with the largest fitness value is regarded as the elite ant lion;

[0123] Step 5: Each ant performs random walks and updates its position according to the current state of the antlion population to achieve global optimization.

[0124] In specific implementation, the process of ants performing full optimization includes:

[0125] Select an antlion for each ant via roulette or other selection mechanism;

[0126] According to the dynamic coefficient a, the ant individual is selected to cross with the ant lion selected by the roulette wheel, or cross with the elite ant lion to update the position of the ant individual; the formula for updating the position of the ant individual is:

[0127]

[0128] In the formula, Ant i (t+1) indicates the position of the ant in the t+1th iteration; i (t) represents the position of the antlion selected by ant individual i in the roulette wheel in the tth iteration; BestAntlion(t) represents the position of the elite antlion in the tth iteration; is a crossover symbol, indicating that the codes of individual A and individual B are cross-operated; rand is a random value in the interval [0,1]; a is a dynamic coefficient that changes with the current number of iterations, and t is the current number of iterations;

[0129] The dynamic coefficient calculation formula is:

[0130]

[0131] Where t is the current iteration number and T is the maximum iteration number.

[0132] By modifying the ants' random walk rule to adapt it to the search of discrete variables, the ants can effectively search in the discrete solution space. This makes the ants' search behavior more consistent with the actual constraints of the scheduling problem and improves the feasibility of the solution. By modifying the ants' random walk rule to a roulette-based selection of antlions and crossover, the algorithm is more suitable for the search problem of discrete variables. This modification enables the algorithm to explore more effectively in the discrete solution space when solving the scheduling problem, improving the algorithm's adaptability to practical problems.

[0133] It can also improve the diversity of solutions. The roulette selection mechanism provides each ant with the opportunity to choose a different antlion, increasing the diversity of individuals in the ant population. At the same time, the introduction of the dynamic coefficient changes the strategy of ants in selecting crossover objects at different iteration stages, further promoting the diversity of solutions in the population. This diversity helps the algorithm to jump out of the local optimal solution and explore a wider solution space. In addition, the setting of the dynamic coefficient α can adjust the selection strategy of ant individuals for crossover individuals during the global search process; to avoid elite antlions falling into the local optimal solution too early in the early iterations. By dynamically adjusting the value of α, the algorithm can balance the search breadth and depth of ants, ensuring that ants can effectively explore a better solution space while maintaining a certain degree of randomness.

[0134] It should be noted that, in the specific implementation, when the ant individual crosses with the ant lion selected by the roulette wheel or the elite ant lion, the machine selection code and the unit selection code use joint uniform crossover, and the process sorting code uses IPOX crossover. (e.g. Figure 4 (shown)

[0135] Such a setting can improve the adaptability of the encoding type. Machine selection encoding and unit selection encoding are usually integer encoding or combination encoding, which often involve a limited discrete selection set. The use of joint uniform crossover can ensure that during the crossover process, the corresponding encoding parts of the two parent individuals can exchange genes in an equal probability manner, thereby maintaining the diversity of solutions and potentially producing better offspring. Process sorting encoding involves the order of processes, which is crucial for scheduling problems. IPOX (Iterated Pairwise Order Crossover) crossover is a crossover operator specifically used for sequence problems. It can effectively maintain the relative order relationship between processes, reduce illegal solutions generated during the crossover process, and thus improve the quality of solutions. It can also enhance the global search ability of the algorithm. The ant lions or elite ant lions selected by roulette are usually individuals with better performance in the current population. Crossing with these individuals helps ant individuals absorb the characteristics of excellent individuals, thereby accelerating the evolution of the entire population. Different types of encoding use different crossover strategies, which can more accurately handle the characteristics of various encoding types and avoid gene destruction or degradation of solution quality due to improper crossover methods. This helps improve the global search capability of the algorithm, enabling it to search for high-quality solutions in a wider solution space.

[0136] In addition, local search and neighborhood exploration can also be promoted. After the crossover operation, a series of local search or neighborhood exploration steps are usually followed to further optimize the quality of the solution. The offspring generated by the hybrid crossover strategy not only inherit the excellent characteristics of the parent individuals, but also introduce new gene combinations through crossover, which provides more possibilities and directions for subsequent local searches. Especially for the case where the process sorting code adopts IPOX crossover, since IPOX crossover can maintain the relative order relationship between the processes, the generated offspring are more likely to maintain the legitimacy of the solution during the local search process, thereby reducing unnecessary repair operations and computational costs. At the same time, since different types of codes use appropriate crossover methods, the algorithm is more stable during the search process and is not easy to fall into the local optimal solution.

[0137] Step 6: Construct the neighborhood solution of each antlion according to the preset method and perform local optimization.

[0138] In specific implementation, the process of constructing the neighborhood solution of each antlion includes:

[0139] For the individual ant lion machine, select the MS code and the unit select the CS code; randomly select 2-3 coding positions, determine the machine number with the minimum processing time in the processing equipment concentration of the coding position process, and replace the original coding position; the CS code is corrected and adjusted as the MS code changes;

[0140] For the process selection OS code of the antlion individual, three process code positions are randomly selected and exchanged with each other to generate 6 sections of neighborhood process selection codes;

[0141] The six neighborhood process selection codes are combined with the adjusted MS code and CS code to obtain six groups of neighborhood solutions.

[0142] (like Figure 5 (shown)

[0143] Compare the fitness values ​​of the antlion individual and its six groups of neighborhood solutions, and take the position with the largest fitness value as the local optimization result of the antlion.

[0144] Such a setting can enhance the local search capability. By randomly selecting 2-3 positions in the machine selection MS code and the unit selection CS code, replacing them based on the principle of optimizing the process processing time, and exchanging the three process positions in the process selection OS code, multiple groups of neighborhood solutions are generated. This method effectively conducts a detailed search in the neighborhood of the ant lion individual. This local search can help the algorithm find a better solution near the discovered high-quality solution, thereby accelerating the convergence process. It can also improve the diversity of solutions. Although the search is conducted in a local range, the random selection of code positions and process exchange methods still maintain the diversity of the search. This diversity helps the algorithm avoid falling into the local optimal solution and increases the possibility of discovering the global optimal solution. When the MS code changes, the CS code can automatically make corrections and adjustments to ensure the legitimacy of the solution. This linkage mechanism ensures the integrity and validity of the solution during the search process and avoids illegal or invalid solutions caused by inconsistent codes. In addition, by constructing the neighborhood solution of the ant lion and performing local optimization, the overall performance of the algorithm has been significantly improved. The strategy of combining local search with global search enables the algorithm to perform detailed search in the local range while maintaining the global search capability, thereby improving the solution quality and efficiency of the algorithm.

[0145] Step 7: Combine the results of Step 5 and Step 6 to update the position of each ant lion and re-determine the elite ant lion.

[0146] Step 8: According to the preset method, perform variable neighborhood search on the re-determined elite ant lion to obtain the final elite ant lion of the current iteration.

[0147] In specific implementation, the process of variable neighborhood search includes:

[0148] Step 8.1: Take the elite ant lion as the initial solution X and set the maximum number of neighborhood searches q max , and set the parameters q=0, k=1;

[0149] Step 8.2: Use the neighborhood structure N for the initial solution X k (k=1,2…,n) performs local search. If the fitness f(X') of the generated new individual X' is greater than the fitness f(X) of the original individual, the original individual is replaced by X' and k=1; otherwise, k=k+1; where N k represents the kth preset neighborhood structure; n is the type of the preset neighborhood structure;

[0150] like Figure 6 As shown, four neighborhood search structures are designed in this embodiment, including three neighborhood structures for OS coding and one neighborhood structure for MS and CS coding.

[0151] Step 8.3: Determine whether k>n. If so, go to step 8.4; otherwise, go to step 8.2;

[0152] Step 8.4: Let q = q + 1, and determine whether the neighborhood search termination condition q>q is satisfied max If it is satisfied, jump to step 8.5; if it is not satisfied, set k=1 and go to step 8.2;

[0153] Step 8.5: The algorithm ends and the result of the neighborhood search is output as the final elite antlion of the current iteration.

[0154] Such a setting can enhance the global search capability. Although variable neighborhood search is a local search technique, by constantly switching the neighborhood structure (N k ) and restart the search from the elite antlion. The algorithm can jump out of the local optimal solution to a certain extent and increase the breadth of the global search. This helps to avoid premature convergence of the algorithm and increase the possibility of finding the global optimal solution. It can also improve the flexibility of the system: by introducing a variety of preset neighborhood structures (N k ), the algorithm can flexibly adjust the search strategy according to the characteristics of the problem and the current search status, enhancing the adaptability and flexibility of the system. This is of great significance for dealing with complex and changeable manufacturing environments. It can also improve the accuracy of the solution. By continuously performing local searches in the neighborhood of the elite ant lion (that is, the current optimal solution), the algorithm can try multiple possible solution space areas, thereby having the opportunity to find a solution that is better than the current optimal solution. This sophisticated search strategy helps to improve the accuracy and performance of the solution. And it can accelerate the convergence speed. Although variable neighborhood search increases the computational complexity of the algorithm, it also accelerates the convergence speed of the algorithm to a certain extent. Because during the search process, the algorithm can more quickly identify and focus on potential optimal solution areas, reducing unnecessary search space exploration.

[0155] Step 9: Determine whether the number of iterations reaches the preset maximum number of iterations maxGen. If not, return to step 5; if reached, take the scheduling plan corresponding to the final elite ant lion of the current iteration as the best scheduling plan.

[0156] The scheduling scheme is represented by three-segment coding, and the maximum fuzzy completion time is obtained by greedy decoding, and then converted into a fitness value by defuzzification technology. This process is simple and efficient. The calculation method of the fitness value (i.e., the inverse of the maximum fuzzy completion time) can intuitively reflect the quality of the solution, which is convenient for the algorithm to select and compare. The algorithm can handle complex cross-unit scheduling problems involving fuzzy time and three-segment coding. By introducing fuzzy number representation and three-segment coding technology, the algorithm can more accurately describe the uncertainty factors in actual production and improve the adaptability and practicality of the scheduling model.

[0157] The antlion optimization algorithm implements global search through the random walk and position update of ants, which can explore a wide area in the solution space and help discover new potential solutions. At the same time, by constructing the neighborhood solution of the antlion and performing local optimization, the algorithm can perform a fine search near the discovered high-quality solution to improve the accuracy of the solution. This combination of global and local search enables the algorithm to quickly converge to a high-quality solution while maintaining search diversity. In addition, the elite strategy ensures that the algorithm continues to evolve towards a better solution during the iteration process by retaining and optimizing the individuals with the highest fitness values ​​(elite antlions). Variable neighborhood search further optimizes the elite antlions, avoids falling into local optimality by exploring different neighborhood structures, and improves the global search capability of the algorithm.

[0158] S5. Use the best scheduling solution obtained in S4 to carry out actual cross-unit production scheduling.

[0159] This method uses triangular fuzzy numbers to characterize the fuzzy processing time and fuzzy transportation time in the workpiece processing process, which can more accurately reflect the uncertainty in actual production. Triangular fuzzy numbers can fully describe the fuzziness and uncertainty range of time variables by defining three parameters: minimum value, most likely value and maximum value. This representation method enables the scheduling model to more truly reflect the actual situation in the production process, thereby improving the accuracy and reliability of the model. The scheduling model takes minimizing the fuzzy maximum completion time as the optimization goal. By introducing fuzzy number representation and fuzzy optimization objectives, the scheduling model can better cope with uncertain factors in the production process (such as equipment failure, raw material delay, etc.) and improve the robustness and adaptability of production scheduling. Since the scheduling model takes into account the uncertainty and multi-objective optimization requirements in the production process, the generated scheduling plan is more in line with the actual production situation, which helps enterprises improve decision-making efficiency and production benefits. Fuzzy time representation and ant lion optimization algorithm are used to make the scheduling plan have a certain adaptability to changes in the external environment and enhance the robustness of the system.

[0160] Considering the discrete optimization characteristics of the cross-unit scheduling problem with fuzzy time, this method performs three-stage coding on each potential scheduling scheme based on process sorting, machine selection and manufacturing unit selection, and then solves it through the ant lion optimization algorithm after corresponding discretization improvement, which can ensure the effectiveness of the obtained scheduling scheme. The three-stage coding decomposes the scheduling problem into three sub-problems: process sorting, machine selection and manufacturing unit selection. By encoding each sub-problem separately, the key information in the scheduling scheme can be more clearly represented, which is convenient for subsequent optimization algorithms to process. In addition, in the ant lion optimization algorithm, the use of three-stage coding can narrow the search space, enabling the algorithm to find high-quality solutions faster. Since the coding of each sub-problem is relatively independent, the algorithm can perform more detailed searches in the solution space of each sub-problem, thereby improving the search efficiency. In addition, the three-stage coding ensures that the generated scheduling scheme meets the constraints in actual production (such as dependencies between processes, capacity limitations of equipment, etc.) through clear coding rules. This coding method enables the algorithm to automatically exclude invalid solutions that do not meet the constraints during the search process, thereby improving the feasibility of the final solution. Moreover, the simplicity and clarity of the three-stage coding makes it easier to implement the algorithm. Developers can design the corresponding algorithm steps and data structures according to the coding rules to achieve efficient scheduling optimization. At the same time, the three-stage coding is also convenient for debugging and verifying the algorithm, because the coding of each sub-problem is relatively independent, so the correctness of the algorithm can be verified by testing each sub-problem separately.

[0161] This method can obtain an effective cross-unit scheduling plan, ensure that it is consistent with the actual processing situation, and can provide effective guidance for actual production.

[0162] Embodiment 2

[0163] In order to verify the effectiveness of the antlion optimization algorithm proposed in this invention, a specific engineering case is used to test it. The example includes 12 workpieces, 10 equipment and three manufacturing units. The equipment information in the manufacturing unit is shown in the following table.

[0164]

[0165] The fuzzy transportation time between manufacturing units is shown in the following table.

[0166]

[0167] The manufacturing unit to which the parts belong and the status of abnormal parts are shown in the following table.

[0168]

[0169]

[0170] Parts process information such as Figure 7 The main parameters of the ant lion optimization algorithm are set as follows: the number of algorithm iterations is maxGen = 100, the ant lion population is P l =50, the ant population is P a =50, the maximum number of neighborhood searches is 10 times.

[0171] Figure 8 This is the optimal fuzzy Gantt chart obtained by the present invention, and its optimal fuzzy maximum completion time is (143, 173, 194) min. In the scheduling chart, the triangle or rectangle below the machine scale is the fuzzy start time of a certain process in the processing of the machine, and the triangle above the machine scale represents the fuzzy completion time of a certain process in the processing of the machine. The process number and the size of the fuzzy time are respectively represented around the vertices of the triangle.

[0172] The total production time of the seven actual manual production schedules in recent production examples are: 153min, 167min, 174min, 183min, 186min, 210min, and 223min respectively. The median is taken as the most likely completion time; the minimum value is the shortest completion time; the maximum value is the latest completion time, that is, assuming that the original actual fuzzy scheduling plan is (153, 183, 223)min. Comparing the optimal fuzzy completion time obtained by the present invention with it, it can be seen that the ant lion optimization algorithm can effectively obtain a better fuzzy cross-unit scheduling solution, and is far superior to the original scheduling solution in terms of the shortest, most likely, and latest time, respectively, verifying the effectiveness of the algorithm of the present invention.

[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A cross-unit fuzzy scheduling method based on the ant lion optimization algorithm, It is characterized in that The following steps are involved: S1. Use triangular fuzzy numbers to characterize the fuzzy processing time and fuzzy transportation time in the workpiece processing process; take minimizing the fuzzy maximum completion time as the optimization goal, and build a scheduling model for cross-unit scheduling production; S2, obtaining the processing data information and cross-unit transportation data information of the workpiece, and generating all possible potential scheduling schemes; the processing data information includes parts, processes, optional equipment and corresponding fuzzy processing time; The cross-unit transportation data information includes fuzzy transportation time; S3, three-stage coding of each potential scheduling scheme based on process sequencing, machine selection, and manufacturing unit selection; S4. Based on the three-stage coding of each potential scheduling solution, the scheduling model is solved using the antlion optimization algorithm to obtain the best scheduling solution; S5. Use the best scheduling solution obtained in S4 to carry out actual cross-unit production scheduling.

2. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm according to claim 1, Features: In S4, the process of solving the scheduling model by the Ant Lion Optimization Algorithm includes: Step 1: Initialize algorithm parameters, including the ant population size P a , the antlion population size P L , the maximum number of iterations of the algorithm maxGen; Step 2: Generate P using random initialization method a +P L Three-part coding of each individual; Step 3: Greedily decode the three-segment code of each individual to obtain the maximum fuzzy completion time and defuzzify it. Then, the reciprocal of the defuzzified value is taken as the fitness value of the individual. Step 4: Sort the individuals according to their fitness values ​​from large to small, and sort the top P L The number of individuals is regarded as the ant lion population, and the remaining P a The number of individuals is regarded as the ant population; and the individual with the largest fitness value is regarded as the elite ant lion; Step 5: Each ant performs random walks and updates its position according to the current state of the antlion population to achieve global optimization; Step 6: Construct the neighborhood solution of each antlion according to the preset method and perform local optimization; Step 7: Combine the results of Step 5 and Step 6 to update the position of each ant lion and re-determine the elite ant lion; Step 8: According to the preset method, the re-determined elite ant lion is subjected to variable neighborhood search to obtain the final elite ant lion of the current iteration; Step 9: Determine whether the number of iterations reaches the preset maximum number of iterations maxGen. If not, return to step 5; if reached, take the scheduling plan corresponding to the final elite ant lion of the current iteration as the best scheduling plan.

3. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm as claimed in claim 2, Features: The expression of the defuzzification operation is: In the formula, Represents triangular fuzzy time The defuzzified value of 1 Indicates the earliest process completion time or the workpiece cross-unit transportation time, a M represents the most likely process completion time or the workpiece cross-unit transportation time, a 2 Indicates the latest operation completion time or the cross-unit transportation time of the workpiece.

4. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm as claimed in claim 2, Features: In step 5, the ant's full search process includes: Select an antlion for each ant via roulette or other selection mechanism; According to the dynamic coefficient a, the ant individual is selected to cross with the ant lion selected by the roulette wheel, or cross with the elite ant lion to update the position of the ant individual; the formula for updating the position of the ant individual is: In the formula, Ant i (t+1) indicates the position of the ant in the t+1th iteration; i (t) represents the position of the antlion selected by ant individual i in the roulette wheel in the tth iteration; BestAntlion(t) represents the position of the elite antlion in the tth iteration; is a crossover symbol, indicating that the codes of individual A and individual B are cross-operated; rand is a random value in the interval [0,1]; a is a dynamic coefficient that changes with the current number of iterations, and t is the current number of iterations; The dynamic coefficient calculation formula is: Where t is the current iteration number and T is the maximum iteration number.

5. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm as claimed in claim 4, Features: When individual ants cross with roulette-selected antlions or elite antlions, the machine selection code and unit selection code use joint uniform crossover, and the process sorting code uses IPOX crossover.

6. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm as claimed in claim 2, Features: In step 6, the process of constructing the neighborhood solution of each antlion includes: For the individual ant lion machine, select the MS code and the unit select the CS code; randomly select 2-3 coding positions, determine the machine number with the minimum processing time in the processing equipment concentration of the coding position process, and replace the original coding position; the CS code is corrected and adjusted as the MS code changes; For the process selection OS code of the antlion individual, three process code positions are randomly selected and exchanged with each other to generate 6 sections of neighborhood process selection codes; The 6-segment neighborhood process selection codes are combined with the adjusted MS code and CS code to obtain 6 groups of neighborhood solutions; Compare the fitness values ​​of the antlion individual and its six groups of neighborhood solutions, and take the position with the largest fitness value as the local optimization result of the antlion.

7. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm as claimed in claim 2, Features: In step 8, the process of variable neighborhood search includes: Step 8.1: Take the elite ant lion as the initial solution X and set the maximum number of neighborhood searches q max , and set the parameters q=0, k=1; Step 8.2: Use the neighborhood structure N for the initial solution X k (k=1,2…,n) performs local search. If the fitness f(X') of the generated new individual X' is greater than the fitness f(X) of the original individual, the original individual is replaced by X' and k=1; otherwise, k=k+1; where N k represents the kth preset neighborhood structure; n is the type of the preset neighborhood structure; Step 8.3: Determine whether k>n. If so, go to step 8.4; otherwise, go to step 8.2; Step 8.4: Let q = q + 1, and determine whether the neighborhood search termination condition q>q is satisfied max If it is satisfied, jump to step 8.5; if it is not satisfied, set k=1 and go to step 8.2; Step 8.5: The algorithm ends and the result of the neighborhood search is output as the final elite antlion of the current iteration.

8. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm according to claim 1, Features: The triangular fuzzy number representing the fuzzy processing / transportation time is In the formula, represents the fuzzy processing / transportation time; x 1 Represents the earliest process / transportation completion time; x M represents the most likely process / transportation completion time; x 2 Represents the latest operation / transport completion time.

9. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm according to claim 1, Features: The objective function of the scheduling model is: In the formula, Indicates workpiece J i The fuzzy completion time of the task; n is the number of workpieces.

10. The cross-unit fuzzy scheduling method based on the ant lion optimization algorithm according to claim 9, Features: The constraints of the scheduling model include: In the formula, Indicates workpiece J i The fuzzy completion time of the j-th process; Indicates workpiece J i The fuzzy start time of the j+1th process; N represents the set of workpiece serial numbers; O i For workpiece J i A set of processes; Indicates workpiece J i The jth process is on machine M k The fuzzy completion time on Indicates workpiece J i The jth process is on machine M k Fuzzy start time on Y ijk When process O ij On device M k 1 when processing, otherwise 0; Indicates workpiece J i The jth process is on machine M k Fuzzy processing time on M ij For workpiece J i The set of processing equipment for the j-th process; Indicates manufacturing unit M k With M k′ The fuzzy transportation time between ijk When process O ij On device M k It is 1 when processing, otherwise it is 0; m represents the total number of processing equipment.

Citation Information

Cited By

  • Flexible workshop scheduling optimization method and system based on improved ant lion optimization algorithm

    CN120297700A