Precise equipment production workshop scheduling method and system considering resource cooperation and equipment reconstruction

Abstract

The invention belongs to the related technical field of workshop scheduling, and discloses a precision equipment production workshop scheduling method and system considering resource cooperation and equipment reconfiguration, and the method comprises the following steps: (1) constructing a precision equipment production workshop scheduling model considering resource cooperation and equipment reconfiguration, and enabling the scheduling model to perform scheduling on the basis of considering the equipment reconfiguration time; meanwhile, production equipment and workers in the workshop are scheduled, and the optimization target is to minimize the maximum completion time; (2) carrying out optimization solution on the scheduling model by adopting a two-stage coevolution algorithm, and obtaining an optimal solution by using a double-layer coding structure and an active decoding scheme considering public idle time; in the first stage, a self-adaptive genetic algorithm is adopted to carry out global search on a scheduling model solution space, and in the second stage, a simulated annealing algorithm of neighborhood disturbance is adopted to carry out local search on elite individuals in the first stage. According to the method, the defect of relatively large scheduling scheme deviation caused by neglecting equipment reconstruction and worker resources in scheduling modeling can be overcome.

Description

Technical Field

[0001] The present invention belongs to the technical field related to workshop scheduling, and more specifically, relates to a method and system for scheduling a precision equipment production workshop taking resource collaboration and equipment reconstruction into consideration. Background Art

[0002] The production process of precision equipment is complex and requires high precision, and has extremely high requirements for production efficiency and product quality, such as high-precision aerospace equipment. Scheduling in precision equipment production workshops involves complex coordination and optimization of multiple processes and multiple resources, and requires efficient completion of production tasks under limited time and resource conditions. In traditional scheduling models, the focus is mainly on the scheduling optimization of a single resource (such as production equipment), while less consideration is given to the collaboration between different resources (such as equipment and workers). However, in the actual production environment of precision equipment, equipment and workers are two interdependent and indispensable important resources. Ignoring the optimization of either one will lead to a decrease in overall production efficiency.

[0003] In addition, existing scheduling methods often fail to take equipment reconstruction time into consideration, resulting in unsatisfactory application effects in the manufacturing process. After completing the previous processing task, the equipment usually needs to replace specific molds, cutting tools or other tools to adapt to the specifications of the next precision equipment. This process is called equipment reconstruction. The molds required for different equipment processing steps are often different, and their reconstruction time is also different. For example, in the manufacturing of high-precision aerospace equipment, some equipment processing requires the installation of special cutting tools on the equipment. However, many scheduling models ignore the key process of equipment reconstruction, cannot accurately reflect the time consumption in actual production, and cannot effectively respond to complex and changing production needs.

[0004] For high-end manufacturing industries such as precision equipment manufacturing, the production process requires high precision and reliability, and traditional scheduling methods cannot meet actual production needs. Therefore, the study of scheduling methods for precision equipment production workshops that consider resource collaboration and equipment reconstruction can make the scheduling plan more in line with actual production conditions, and has important practical application value and broad promotion prospects in improving production efficiency, shortening production cycles, and improving economic benefits. Summary of the invention

[0005] In response to the above defects or improvement needs of the prior art, the present invention provides a scheduling method and system for a precision equipment production workshop that takes into account resource collaboration and equipment reconstruction, which aims to solve the problem of large scheduling plan deviations caused by ignoring equipment reconstruction and worker resources in existing production scheduling modeling.

[0006] To achieve the above object, according to one aspect of the present invention, a scheduling method for a precision equipment production workshop considering resource collaboration and equipment reconstruction is provided, characterized in that the scheduling method comprises the following steps:

[0007] Step 1: Build a scheduling model for a precision equipment production workshop that considers resource collaboration and equipment reconstruction. The scheduling model schedules the production equipment and workers in the workshop while considering the equipment reconstruction time. The optimization goal is to minimize the maximum completion time.

[0008] Step 2: A two-stage co-evolutionary algorithm is used to optimize and solve the scheduling model. A two-layer coding structure is used to represent the feasible solution of the scheduling model. An active decoding scheme that takes into account the common idle time is used to obtain the individual objective function value. In the first stage, an adaptive genetic algorithm is used to perform a global search of the solution space of the scheduling model. In the second stage, a simulated annealing algorithm with neighborhood perturbation is used to perform a local search for the elite individuals in the first stage.

[0009] Furthermore, the process O in the scheduling model ij The calculation formula for the time to start processing is:

[0010]

[0011] In the formula, i represents the index of precision equipment, ranging from i=1, 2, ..., n, where n is the total number of equipment; j represents the process index of equipment i, ranging from j=1, 2, ..., J i , J i is the total number of processes of precision equipment i; k is the index of the equipment, ranging from k = 1, 2, ..., m, where m is the total number of equipment; e is the index of the worker, ranging from e = 1, 2, ..., w, where w is the total number of workers; g has the same meaning as i; h has the same meaning as j; E k represents the current idle time of device k, that is, the end time of the current task on device k; p' ijk Indicates that equipment k has not yet processed other processes and directly processes process O ij Required setup time; p ghijk Indicates that equipment k processes O first gh Reprocessing ij Reconstruction time; C ij Indicates process O ij The time to finish processing; F e Represents the idle time of worker e, that is, the time it takes for worker e to complete the reconstruction task of the last production equipment.

[0012] Furthermore, the maximum completion time is the maximum value of the processing time of the last process of all precision equipment on the equipment. The calculation formula of the maximum completion time in the scheduling model is:

[0013]

[0014] Furthermore, the first layer of the two-layer coding structure is the equipment allocation layer, which represents the equipment allocated to each process during processing; the second layer is the process sorting layer, which represents the sequence of all processing tasks; workers are allocated according to the equipment allocation results of the process and the process sorting results: idle workers are allocated first. If there are no idle workers, the process processing needs to wait for idle workers to appear before it can proceed.

[0015] Furthermore, an active decoding scheme is adopted that takes into account the common idle time of equipment and workers: ij Arranged to be processed on equipment k, O ij The processing time on equipment k is expressed as p ijk Indicates that the idle time interval of device k is [Lb k ,Ub k ] indicates that if the idle time interval can be allocated to O ij , the following conditions must be met:

[0016]

[0017] If this condition is met, process O ij The calculation formula for the start time is updated to:

[0018]

[0019] Furthermore, the maximum number of iterations of the two-stage co-evolutionary algorithm is set to G max ; The maximum crossover rate is Pc max , the minimum crossover rate is Pc min , the maximum mutation rate is Pm max , the minimum mutation rate is Pm min , the initial temperature is T start , annealing temperature is T end , the number of local search iterations is L max , the annealing rate is λ, and the population size is NP; individuals are generated by random initialization method, and each individual in the population adopts a two-layer coding structure to represent the feasible solution of the above scheduling model, and the active decoding scheme considering the common idle time is adopted to obtain the individual objective function value.

[0020] Furthermore, step 2 includes the following sub-steps:

[0021] Step 1: Set the algorithm parameters and use a two-layer coding structure to represent the feasible solution of the scheduling model;

[0022] Step 2, initialize the population individuals, and use the active decoding scheme considering the public idle time to obtain the individual objective function value;

[0023] Step 3, determine the ranking of each individual in the population, and calculate the adaptive crossover rate and mutation rate;

[0024] Step 4, perform selection operation, mutation operation and crossover operation in sequence;

[0025] Step 5, update the population and select elite individual chromosomes;

[0026] Step 6: Perform neighborhood perturbation on each elite individual. If a better solution is obtained, update the elite individual of the population; otherwise, execute the Metropolis criterion to determine whether to accept the new solution;

[0027] Step 7. If the maximum number of disturbances L is reached max , then jump to Step8; otherwise jump to Step6 and repeatedly perform neighborhood perturbations on the current solution until the termination condition of the neighborhood perturbation is reached;

[0028] Step 8, if the annealing temperature T is reached end , then jump to Step 9; otherwise, reset the annealing rate λ, that is, randomly generate a random number between [0,1] as the new annealing rate; and jump to Step 6, repeatedly perform local search on the elite individuals until the annealing temperature is reached;

[0029] Step 9: If the current number of iterations of the population is equal to the maximum number of iterations G max , then output the current optimal solution; otherwise jump to step Step3 and iterate repeatedly until the termination condition is met.

[0030] Furthermore, the adaptive crossover rate Pc of individual i in the population i and adaptive mutation rate Pm i According to its fitness fit i And the average fitness fit mean The corresponding calculation formula is:

[0031]

[0032] In the formula, fit max Represents the maximum fitness value of the population, and the fitness value is the same as the optimization target value.

[0033] The present invention also provides a precision equipment production workshop scheduling system that takes resource collaboration and equipment reconstruction into consideration. The system includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, it executes the precision equipment production workshop scheduling method that takes resource collaboration and equipment reconstruction into consideration as described above.

[0034] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the precision equipment production workshop scheduling method that considers resource collaboration and equipment reconstruction as described above.

[0035] In general, compared with the prior art, the above technical solutions conceived by the present invention mainly have the following beneficial effects:

[0036] 1. According to the production characteristics of precision equipment, a scheduling model for the precision equipment production workshop is established which takes into account the resource collaboration relationship and equipment reconstruction time. The equipment reconstruction and the equipment processing tasks are considered simultaneously. This can truly and clearly describe the reconstruction process of the production equipment in the workshop processing different precision equipment, and can reflect the collaborative relationship between the two limited resources of equipment and workers. This can overcome the defect of large scheduling plan deviation caused by ignoring equipment reconstruction and worker resources in production scheduling modeling.

[0037] 2. The double-layer coding structure based on equipment allocation and process sorting reduces the coding of the worker allocation layer, which can significantly reduce the length of individual chromosomes, reduce the algorithm encoding and decoding pressure, and reduce the difficulty of executing crossover operations and mutation operations. The active decoding scheme considering the common idle time ensures the efficiency and optimality of the decoding process. Since the optimization goal is to minimize the maximum completion time, the active decoding scheme can obtain a better solution for the corresponding scheduling structure under the coding space and improve the search efficiency.

[0038] 3. The adaptive genetic algorithm in the first stage can effectively traverse the solution space of the precision equipment production scheduling model. The adaptive crossover rate and mutation rate can prevent the population from falling into the local optimal solution and provide high-quality elite chromosomes for the local search in the second stage. The simulated annealing algorithm with domain perturbation in the second stage can optimize the allocation of equipment and workers to the greatest extent based on the task sequence given in the first stage, so as to guide the evolution of the population and improve the convergence speed of the algorithm.

[0039] 4. A two-stage co-evolutionary algorithm is used to optimize the scheduling model of the production workshop of precision equipment. It can traverse more feasible solutions to ensure global optimality. The scheduling results are in line with production reality and can be directly applied to the production of precision equipment production workshops. It can also greatly improve the utilization rate of equipment resources and worker resources, thereby improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flow chart of a precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction provided by the present invention;

[0041] Figure 2 (a) and (b) are schematic diagrams of the dual-layer structure of the scheduling model and the active decoding scheme considering the common idle time of equipment and workers, respectively;

[0042] Figure 3 yes Figure 2 Gantt chart after active decoding of chromosomes in the double-layer coding structure;

[0043] Figure 4 Schematic diagram of two neighborhood disturbances based on critical paths of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0045] See also Figure 1 and Figure 2 The present invention provides a scheduling method for a precision equipment production workshop that takes resource collaboration and equipment reconstruction into consideration. The scheduling method constructs a scheduling model for a precision equipment production workshop that includes two types of resources, equipment and workers, under the premise of considering the equipment reconstruction time, and uses a two-stage co-evolutionary algorithm to optimize and solve the scheduling model to obtain the optimal scheduling plan for equipment resources and worker resources in the workshop, which can significantly improve production efficiency, shorten production cycle, and thereby increase economic benefits.

[0046] Specifically, the method first needs to establish a scheduling model for a precision equipment production workshop that takes into account resource collaboration and equipment reconstruction time, and then uses a two-stage collaborative evolutionary algorithm to optimize and solve the scheduling model. At the same time, a double-layer coding structure and an active decoding scheme that takes into account common idle time are used to obtain the optimal solution for the population; among them, in the first stage, an adaptive genetic algorithm is used to perform a global search of the solution space of the scheduling model, and in the second stage, a simulated annealing algorithm with neighborhood perturbation is used to perform a local search for the elite individuals in the first stage.

[0047] Specifically, the scheduling method mainly includes the following steps:

[0048] Step 1: Build a scheduling model for a precision equipment production workshop that takes resource collaboration and equipment reconstruction into consideration. The scheduling model schedules the production equipment and workers in the workshop while taking the equipment reconstruction time into consideration. The optimization goal is to minimize the maximum completion time.

[0049] The basic description of the scheduling model is: n precision equipment needs to be processed in a workshop containing m production equipment, and each precision equipment contains J i Process O ij , the set of processable equipment and the corresponding processing time for each process are known. After processing a process, the equipment needs to replace the mold to continue processing the next process. This process is called equipment reconstruction. The reconstruction time of the equipment between different sequences is known. At the same time, there are w workers in the workshop, and each equipment reconstruction requires the allocation of workers. If there are no idle workers at present, the equipment enters the waiting state. At time 0, all equipment and workers can perform corresponding production tasks, and all processing tasks are performed on the equipment.

[0050] Process O in the scheduling model ij The calculation formula for the time to start processing is:

[0051]

[0052] In the formula, i represents the index of precision equipment, ranging from i=1, 2, ..., n, where n is the total number of equipment; j represents the process index of equipment i, ranging from j=1, 2, ..., J i , J i is the total number of processes of precision equipment i; k is the index of the equipment, ranging from k = 1, 2, ..., m, where m is the total number of equipment; e is the index of the worker, ranging from e = 1, 2, ..., w, where w is the total number of workers; g has the same meaning as i; h has the same meaning as j; E k represents the current idle time of device k, that is, the end time of the current task on device k; p' ijk Indicates that equipment k has not yet processed other processes and directly processes process O ij Required setup time; p ghijk Indicates that equipment k processes O first gh Reprocessing ij Reconstruction time; C ij Indicates process O ij Time to finish processing; F e Represents the idle time of worker e, that is, the time it takes for worker e to complete the reconstruction task of the last production equipment.

[0053] It can be obtained that the maximum completion time is the maximum value of the processing time of the last process of all precision equipment on the equipment. The calculation formula of the maximum completion time in the scheduling model is:

[0054]

[0055] Step 2: A two-stage co-evolutionary algorithm is used to optimize and solve the scheduling model. A two-layer coding structure is used to represent the feasible solution of the scheduling model. An active decoding scheme that takes into account the common idle time is used to obtain the individual objective function value. In the first stage, an adaptive genetic algorithm is used to perform a global search of the solution space of the scheduling model. In the second stage, a simulated annealing algorithm with neighborhood perturbation is used to perform a local search for the elite individuals in the first stage.

[0056] A two-layer coding scheme is adopted: the first layer is the equipment allocation layer, which represents the equipment allocated to each process during processing; the second layer is the process sorting layer, which represents the sequence of all processing tasks. Worker resources are allocated based on the first idle, first served principle, reducing the coding of the worker allocation layer;

[0057] An active decoding scheme that considers the common idle time of equipment and workers is adopted: when process O ij Arranged to be processed on equipment k, O ij The processing time on equipment k is expressed as p ijk Indicates that the idle time interval of device k is [Lb k ,Ub k ] indicates that if the idle time interval can be allocated to O ij , the following conditions must be met:

[0058]

[0059] If this condition is met, process O ij The calculation formula for the start time is updated to:

[0060]

[0061] In this implementation, step 2 mainly includes the following sub-steps:

[0062] Step 1. Set the algorithm parameters and use a two-layer coding structure to represent the feasible solution of the scheduling model.

[0063] The maximum number of iterations of the two-stage co-evolutionary algorithm is set to G max ; The maximum crossover rate is set to Pc max , the minimum crossover rate is set to Pc min , the maximum mutation rate is set to Pm max , the minimum mutation rate is set to Pm min , the initial temperature is set to T start , the annealing temperature is set to T end , the number of local search iterations is set to L max, the annealing rate is set to λ, and the population size is set to NP; each individual in the population uses a two-layer coding structure to represent the feasible solution of the above scheduling model. Specifically: the first layer of the two-layer coding structure is the equipment allocation layer, which represents the equipment allocated to each process during processing; the second layer is the process sorting layer, which represents the sequence of all processing tasks; workers are allocated according to the equipment allocation results and the process sorting results of the process: idle workers are allocated first. If there are no idle workers, the process processing needs to wait for idle workers to appear before proceeding.

[0064] Step 2: Initialize the individuals in the population and use an active decoding scheme that takes into account the common idle time to obtain the individual objective function value.

[0065] The random initialization method is used to generate population individuals. Each individual in the population has a double-layer coding structure as described above, and an active decoding scheme is adopted that takes into account the common idle time of equipment and workers. Specifically: ij The processing time on equipment k is expressed as p ijk Indicates that the time interval of device k is [Lb k ,Ub k ] indicates that if the time interval can be allocated to O ij , the following conditions must be met:

[0066]

[0067] If the above conditions are met, process O ij The calculation formula for the start time is updated to:

[0068]

[0069] Step 3: Determine the ranking of each individual in the population and calculate the adaptive crossover rate and mutation rate.

[0070] Arrange all individuals in ascending order according to the optimization target value to determine the ranking of each individual in the population. The higher the ranking, the better the individual. Suppose there are individuals x1, x2, x3, x4 and x5 in the population, and their objective function values ​​are 16, 12, 14, 18 and 15 respectively. The results of ascending order are 12, 14, 15, 16, 18, and the corresponding arrangement of individuals in the population is: x2, x3, x5, x1, x4, that is, x2 is better than other individuals, and x4 is not as good as other individuals. Then, according to its fitness fit i And the average fitness fit mean The proportion of population individual i is adjusted to adjust the adaptive crossover rate Pc i and adaptive mutation rate Pm i , the specific calculation formula is:

[0071]

[0072] From the above formula, we can see that if the individual fitness is higher than the average level, the crossover rate and mutation rate take smaller values, which is conducive to retaining excellent genes; if the individual fitness is lower than the average level, the crossover rate and mutation rate take larger values, increasing the number of gene updates, which is conducive to traversing more feasible solutions.

[0073] Step 4: Perform selection operation, mutation operation and crossover operation in sequence.

[0074] In this step, the first-stage adaptive genetic algorithm is used to perform selection, crossover and mutation operations on individuals in the population. The following are implementation methods for the three operations.

[0075] Selection operation: Use the role model selection method to define the individual with a higher ranking in the population as the role model of individual i, and randomly select an individual from the role model set of each individual to obtain a new population. Specifically: Continue to use the above assumptions, and now the arrangement of individuals in the population is x2, x3, x5, x1, x4. Then the role model individual of x2 and x3 can only be x2, the role model individual of x5 is x2 and x3, the role model individual set of x1 is {x2, x3, x5}, and the role model individual set of x4 is {x2, x3, x5, x1}. Randomly select an individual from the role model set of each individual in turn, assuming it is x2, x2, x3, x3 and x5, then they constitute a new population; it can be found that the new population obtained by the role model selection method may contain the same individuals.

[0076] Crossover operation: select two individuals from the population, perform multi-point crossover on the equipment allocation layer sequence, and perform POX crossover on the process sorting layer sequence. Specifically: assume that individuals 1 and 2 are crossed, if the equipment allocation layer sequence of individual 1 is [1 22 3 1 2 3 1], and the equipment allocation layer sequence of individual 2 is [2 1 2 3 2 2 3 1], and the randomly selected positions are 2 and 5, then after the crossover, the equipment allocation layer sequence of individual 1 is [1 1 2 3 2 2 3 1], and the equipment allocation layer sequence of individual 2 is [2 2 23 1 2 3 1]; if the process sorting layer sequence of individual 1 is [2 1 4 2 3 1 4 3], and the process sorting layer sequence of individual 2 is [4 3 1 2 34 1 2], randomly divide the precision equipment into two sets {1, 3} and {2, 4}, copy the processes of individual 1 belonging to the precision equipment set {1, 3} to C1, fill the vacant positions of C1 with the processes of individual 2 belonging to the precision equipment set {2, 4}, copy the processes of individual 2 belonging to the precision equipment set {1, 3} to C2, and fill the vacant positions of C2 with the processes of individual 1 belonging to the precision equipment set {2, 4}. Then C1 is the process sorting layer sequence [4 1 2 4 3 1 2 3] of individual 1 after crossing, and C2 is the process sorting layer sequence [2 3 1 4 3 2 1 4] of individual 2 after crossing.

[0077] Mutation operation: randomly select r positions for the equipment allocation layer of the current individual, and change the equipment at each position to the minimum load equipment in the optional equipment set of the corresponding process; randomly select r positions for the process sorting layer of the current individual, and select the optimal feasible solution under all its arrangements. Specifically: assuming that the equipment allocation layer sequence of the current individual is [2 3 1 2 2 3 31], and the randomly selected positions are 3 and 6, if the equipment with the smallest load in the optional equipment set of the process corresponding to position 3 is equipment 2, and the equipment with the smallest load in the optional equipment set of the process corresponding to position 6 is equipment 1, then the equipment allocation layer sequence of the current individual after mutation is [2 3 2 2 2 1 3 1]; assuming that the process sorting layer sequence of the current individual is [2 3 1 2 1 4 4 3], and the randomly selected positions are 1, 3, and 6, then there are 3×2×1=6 sequences under all permutations of this position, namely [1 3 2 2 1 4 4 3], [1 3 4 2 1 2 4 3], [2 3 1 2 1 4 4 3], [2 3 4 2 1 2 4 3], [4 3 1 2 1 2 4 3], and [4 32 2 1 1 4 3], where the sequence that matches the equipment allocation layer sequence to obtain the optimal objective function value is the process ranking layer sequence after the current individual variation operation.

[0078] Step 5. Update the population and select elite individual chromosomes.

[0079] The new population after the selection, crossover and mutation operations is re-evaluated, and the top 10% of the individuals in the population are selected as elite individuals according to the new optimization target value, and enter the second stage of local search.

[0080] Step 6: Perform neighborhood perturbation on each elite individual. If a better solution is obtained, update the elite individual of the population; otherwise, execute the Metropolis criterion to determine whether to accept the new solution.

[0081] In this step, the simulated annealing algorithm of the second-stage neighborhood perturbation is used to perform neighborhood perturbation based on the critical path on the elite individuals in the first stage: a random number between [0,1] is randomly generated. If the random number is less than 0.5, the neighborhood perturbation of the same device is performed on the current elite solution; if the random number is greater than or equal to 0.5, the neighborhood perturbation of changing the device is performed on the current elite solution. Specifically:

[0082] Neighborhood disturbance of the same equipment: define the maximum process sequence contained in the critical path on the same equipment as a key block, and exchange the two processes at the beginning and the two processes at the end of each key block. If the two processes at the beginning or the two processes at the end of the block belong to the same precision equipment, they will not be exchanged.

[0083] Neighborhood disturbance of changing equipment: randomly select a process on the critical path that contains two or more optional equipment, delete the process from the processing task of the current equipment, randomly select one from the remaining optional equipment, and insert it into the processing task of the selected equipment.

[0084] If the fitness of the neighborhood solution is fit new Than the fitness of the current solution fit i If the new solution is better, the neighborhood solution is used to replace the current solution. Otherwise, the Metropolis criterion is implemented. If the new solution is accepted, the following conditions must be met:

[0085]

[0086] Where ε is a random number between [0,1], and T is the temperature of the current wheel.

[0087] Step 7: End of neighborhood disturbance.

[0088] If the maximum number of disturbances L is reached max , then jump to Step 8; otherwise jump to Step 6 and repeatedly perform neighborhood perturbations on the current solution until the termination condition of the neighborhood perturbation is met.

[0089] Step 8. Exit the local search process.

[0090] If the annealing temperature T end, then jump to Step 9; otherwise, reset the annealing rate λ, that is, randomly generate a random number between [0,1] as the new annealing rate; and jump to Step 6 to repeat the local search for the elite individuals until the annealing temperature is reached.

[0091] Step 9: Termination of the two-stage co-evolution algorithm.

[0092] If the current iteration number of the population is equal to the maximum iteration number G max , then output the current optimal solution; otherwise jump to step Step3 and iterate repeatedly until the termination condition is met.

[0093] The present invention is further described in detail below with reference to specific embodiments.

[0094] An example with 4 precision equipment to be processed, 3 production equipment and 2 workers is provided: Table 1 shows the processing time of each process on different equipment. No number means that the equipment cannot process this process; Table 2, Table 3 and Table 4 show the reconstruction time on the 1st, 2nd and 3rd equipment respectively; the row name indicates the process to be processed first, the column name indicates the process to be processed later, the start column indicates that the equipment has not processed any process yet, and the process O ij As the preparation time required for the first processing step. For example, equipment 1 first processes O 12 , reprocessing O 21 , the time required for the reconstruction process is 2; Equipment 2 first processes O 22 , reprocessing O 12 , no reconstruction is required.

[0095] Table 1 Processing time information

[0096] <![CDATA[M1]]> <![CDATA[M2]]> <![CDATA[M3]]> <![CDATA[O 11 ]]> 3 5 4 <![CDATA[O 12 ]]> 4 2 3 <![CDATA[O 21 ]]> 2 3 1 <![CDATA[O 22 ]]> 3 4 3 <![CDATA[O 31 ]]> 3 4 5 <![CDATA[O 32 ]]> 3 - 3 <![CDATA[O 41 ]]> 5 3 - <![CDATA[O 42 ]]> 4 3 2

[0097] Table 2 Reconstruction time on device 1

[0098] start <![CDATA[O 11 ]]> <![CDATA[O 12 ]]> <![CDATA[O 21 ]]> <![CDATA[O 22 ]]> <![CDATA[O 31 ]]> <![CDATA[O 32 ]]> <![CDATA[O 41 ]]> <![CDATA[O 42 ]]> <![CDATA[O 11 ]]> 1 0 1 2 1 1 2 2 1 <![CDATA[O 12 ]]> 1 0 0 2 2 2 1 1 0 <![CDATA[O 21 ]]> 1 1 2 0 1 1 1 1 1 <![CDATA[O 22 ]]> 2 1 2 0 0 2 2 2 0 <![CDATA[O 31 ]]> 1 1 2 1 1 0 1 0 1 <![CDATA[O 32 ]]> 1 2 2 2 1 0 0 1 1 <![CDATA[O 41 ]]> 2 0 1 2 2 1 2 0 1 <![CDATA[O 42 ]]> 1 1 2 1 0 1 0 2 0

[0099] Table 3 Reconstruction time on device 2

[0100] start <![CDATA[O 11 ]]> <![CDATA[O 12 ]]> <![CDATA[O 21 ]]> <![CDATA[O 22 ]]> <![CDATA[O 31 ]]> <![CDATA[O 32 ]]> <![CDATA[O 41 ]]> <![CDATA[O 42 ]]> <![CDATA[O 11 ]]> 2 0 1 1 2 1 - 2 0 <![CDATA[O 12 ]]> 1 1 0 2 2 1 - 1 1 <![CDATA[O 21 ]]> 1 0 1 0 0 1 - 0 2 <![CDATA[O 22 ]]> 2 1 0 1 0 2 - 2 1 <![CDATA[O 31 ]]> 1 1 2 2 2 0 - 1 1 <![CDATA[O 32 ]]> - - - - - - - - - <![CDATA[O 41 ]]> 1 0 1 1 1 2 - 0 0 <![CDATA[O 42 ]]> 2 0 1 2 0 2 - 2 0

[0101] Table 4 Reconstruction time on device 3

[0102]

[0103]

[0104] In this example, there are 4 precision equipment to be processed, and the number of processes for each precision equipment is 2. The length of each layer of code in the double-layer coding scheme is 2+2+2+2=8, and the number of optional production equipment for each process is [3, 3, 3, 3, 3, 2, 2, 3]. The number of workers in the workshop is 2, and each worker can operate each equipment for reconstruction.

[0105] See also Figure 1 , this implementation example includes the following steps:

[0106] Step 1: Build a scheduling model.

[0107] According to the equipment processing information and reconstruction time of the instance, a scheduling model for the precision equipment production workshop is constructed which takes into account resource collaboration and equipment reconstruction time. The optimization goal is to maximize the completion time of three devices processing four precision equipment.

[0108] Step 2: Use a two-stage co-evolutionary algorithm to optimize and solve the scheduling model.

[0109] This step 2 mainly includes the following sub-steps:

[0110] Step 1. Set the algorithm parameters and use a two-layer coding scheme to represent the feasible solution of the scheduling model.

[0111] The maximum number of iterations of the two-stage co-evolutionary algorithm is set to G max =200; the maximum crossover rate is set to Pc max =0.9, the minimum crossover rate is set to Pc min =0.5, the maximum mutation rate is set to Pm max =0.1, the minimum mutation rate is set to Pm min =0.01, the initial temperature is set to T start =1000, annealing temperature is set to T end =100, the number of local search iterations is set to L max =10, the annealing rate is set to λ=rand(0,1), and the population size is set to NP=50; each individual in the population uses a two-layer coding structure to represent the feasible solution of the scheduling model. Figure 2 (a) in FIG. 1 shows a double-layer coding structure in this example.

[0112] Step 2: Initialize the population and actively decode to obtain the individual objective function value.

[0113] The random initialization method is used to generate population individuals; the active decoding process in this instance is as follows Figure 2 As shown in (b): Process O ijis scheduled to be processed on device k, searches for the idle time period of device k, and determines whether there are idle workers in this time interval. If so, the idle time period is allocated to O ij , otherwise O ij Arranged at the end of the processing task of equipment k, Figure 3 Shown Figure 2 Gantt chart of the coded sequence in (a) after active decoding.

[0114] Step 3: Determine the ranking of each individual in the population and calculate the adaptive crossover rate and mutation rate.

[0115] Arrange the 50 individuals in ascending order according to the optimization target value to determine the ranking of each individual. The higher the ranking, the better the individual. Individuals with the same optimization target value are ranked first according to the smaller individual number value, that is, x i The smaller the i value, the better the ranking is assigned to ensure that individual ranking values ​​are not repeated.

[0116] Next, according to the individual fitness fit i And the average fitness fit mean The proportion of population individual i is adjusted to adjust the adaptive crossover rate Pc i and adaptive mutation rate Pm i In a certain iteration, the fitness values ​​of 50 individuals are [28 20 21 2521 18 19 17 17 26 30 22 21 23 25 21 17 23 19 19 2219 22 17 22 20 14 25 18 2317 22 25 16 20 22 26 15 22 19 20 19 19 24 21 15 2616 21 22], then the maximum fitness is fit max =30, average fitness fit mean =20.82, further, the crossover rate and mutation rate of each individual can be calculated. Taking the 8th individual as an example, its crossover rate in the current round is The mutation rate is

[0117] Step 4: Perform selection operation, mutation operation and crossover operation in sequence.

[0118] Selection operation: Use the role model selection method to define individuals with higher rankings than individual i in the population as role models for individual i. The role model of the first-ranked individual can only be itself. Randomly select an individual from the role model set of each individual to obtain a new population, which still has 50 individuals.

[0119] Crossover operation: randomly generate a random number R1 between [0,1], if R1≤Pci , then a spouse individual is randomly selected for the current individual, a multi-point crossover operation is performed on the equipment allocation sequence of the two individuals, and a POX crossover operation is performed on the process sorting sequence.

[0120] Mutation operation: Randomly generate a random number R2 between [0,1], if R2≤Pm i , then mutation operations are performed on the equipment allocation sequence and process sorting sequence of the current individual respectively.

[0121] Step 5. Update the population and select elite individual chromosomes.

[0122] The new population individuals after selection, crossover and mutation are decoded to obtain the optimization target values ​​of 50 individuals, and the population is updated; the top 10% individuals with fitness values ​​from small to large in the current population are selected as elite individuals, that is, 5 elite individuals enter the local search stage of the second stage.

[0123] Step 6: Perform neighborhood perturbation on each elite individual. If a better solution is obtained, update the elite individual of the population; otherwise, execute the Metropolis criterion to determine whether to accept the new solution.

[0124] The simulated annealing algorithm of the second-stage neighborhood perturbation is used to perturb the neighborhood of the five elite individuals. If the generated random number is less than 0.5, the neighborhood perturbation of the same device is performed on the current solution. The specific operation is as follows: Figure 4 As shown in (a) in ; if it is greater than or equal to 0.5, the neighborhood perturbation of the current solution is performed by switching the device. The specific operation is as follows Figure 4 As shown in (b) in the figure. If the fitness of the neighborhood solution is fit new Than the fitness of the current solution fit i If the new solution is better, the neighborhood solution is used to replace the current solution. Otherwise, the Metropolis criterion is used to determine whether to accept the new solution.

[0125] Step 7: If the maximum number of perturbations reaches 10, jump to step 8; otherwise, jump to step 6 and repeatedly perform neighborhood perturbations on the current solution until the termination condition of the neighborhood perturbation is reached.

[0126] Step 8. If the annealing temperature reaches 100, jump to step 9; otherwise, randomly generate a random number as the new annealing rate and jump to step 6 to repeat the local search for the elite individuals until the annealing temperature is reached.

[0127] Step 7: If the number of iterations of the two-stage co-evolution algorithm reaches 200, the current optimal solution is output; otherwise, jump to step 3 and iterate repeatedly until the termination condition is met.

[0128] The present invention also provides a precision equipment production workshop scheduling system that takes resource collaboration and equipment reconstruction into consideration. The system includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, it executes the precision equipment production workshop scheduling method that takes resource collaboration and equipment reconstruction into consideration as described above.

[0129] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the precision equipment production workshop scheduling method that considers resource collaboration and equipment reconstruction as described above.

[0130] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A scheduling method for precision equipment production workshop considering resource collaboration and equipment reconstruction, Features: The scheduling method includes the following steps: Step 1: Build a scheduling model for a precision equipment production workshop that considers resource collaboration and equipment reconstruction. The scheduling model schedules the production equipment and workers in the workshop while considering the equipment reconstruction time. The optimization goal is to minimize the maximum completion time. Step 2: A two-stage co-evolutionary algorithm is used to optimize and solve the scheduling model. A two-layer coding structure is used to represent the feasible solution of the scheduling model. An active decoding scheme that takes into account the common idle time is used to obtain the individual objective function value. In the first stage, an adaptive genetic algorithm is used to perform a global search of the solution space of the scheduling model. In the second stage, a simulated annealing algorithm with neighborhood perturbation is used to perform a local search for the elite individuals in the first stage.

2. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 1, Features: Process O in the scheduling model ij The calculation formula for the time to start processing is: In the formula, i represents the index of precision equipment, ranging from i=1, 2, ..., n, where n is the total number of equipment; j represents the process index of equipment i, ranging from j=1, 2, ..., J i , J i is the total number of processes of precision equipment i; k is the index of the equipment, ranging from k = 1, 2, ..., m, where m is the total number of equipment; e is the index of the worker, ranging from e = 1, 2, ..., w, where w is the total number of workers; g has the same meaning as i; h has the same meaning as j; E k represents the current idle time of device k, that is, the end time of the current task on device k; p' ijk Indicates that equipment k has not yet processed other processes and directly processes process O ij Required setup time; p ghijk Indicates that equipment k processes O first gh Reprocessing ij Reconstruction time; C ij Indicates process O ij Time to finish processing; F e Represents the idle time of worker e, that is, the time it takes for worker e to complete the reconstruction task of the last production equipment.

3. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 2, Features: The maximum completion time is the maximum value of the processing time of the last process of all precision equipment on the equipment. The calculation formula of the maximum completion time in the scheduling model is:

4. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 1, Features: The first layer of the two-layer coding structure is the equipment allocation layer, which represents the equipment allocated to each process during processing; the second layer is the process sorting layer, which represents the sequence of all processing tasks; workers are allocated according to the equipment allocation results and the process sorting results of the process: idle workers are allocated first. If there are no idle workers, the process processing needs to wait for idle workers to appear before it can proceed.

5. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 2, Features: An active decoding scheme that considers the common idle time of equipment and workers is adopted: when process O ij Arranged to be processed on equipment k, O ij The processing time on equipment k is expressed as p ijk The idle time interval of device k is expressed as [Lb k ,Ub k ] indicates that if the idle time interval can be allocated to O ij , the following conditions must be met: If this condition is met, process O ij The calculation formula for the start time is updated to:

6. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 5, Features: Set the maximum number of iterations of the two-stage co-evolutionary algorithm to G max ; The maximum crossover rate is Pc max , the minimum crossover rate is Pc min , the maximum mutation rate is Pm max , the minimum mutation rate is Pm min , the initial temperature is T start , annealing temperature is T end , the number of local search iterations is L max , the annealing rate is λ, and the population size is NP; individuals are generated by random initialization method, and each individual in the population adopts a two-layer coding structure to represent the feasible solution of the above scheduling model, and the active decoding scheme considering the common idle time is adopted to obtain the individual objective function value.

7. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 6, Features: Step 2 includes the following sub-steps: Step 1: Set the algorithm parameters and use a two-layer coding structure to represent the feasible solution of the scheduling model; Step 2, initialize the population individuals, and use the active decoding scheme considering the public idle time to obtain the individual objective function value; Step 3, determine the ranking of each individual in the population, and calculate the adaptive crossover rate and mutation rate; Step 4, perform selection operation, mutation operation and crossover operation in sequence; Step 5, update the population and select elite individual chromosomes; Step 6: Perform neighborhood perturbation on each elite individual. If a better solution is obtained, update the elite individual of the population; otherwise, execute the Metropolis criterion to determine whether to accept the new solution; Step 7. If the maximum number of disturbances L is reached max , then jump to Step8; otherwise jump to Step6 and repeatedly perform neighborhood perturbations on the current solution until the termination condition of the neighborhood perturbation is reached; Step 8, if the annealing temperature T is reached end , then jump to Step 9; otherwise, reset the annealing rate λ, that is, randomly generate a random number between [0,1] as the new annealing rate; and jump to Step 6, repeatedly perform local search on the elite individuals until the annealing temperature is reached; Step 9: If the current number of iterations of the population is equal to the maximum number of iterations G max , then output the current optimal solution; Otherwise, jump to Step 3 and iterate repeatedly until the termination condition is met.

8. The precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as claimed in claim 7, Features: Adaptive crossover rate Pc of individual i in the population i and adaptive mutation rate Pm i According to its fitness fit i And the average fitness fit mean The corresponding calculation formula is: In the formula, fit max Represents the maximum fitness value of the population, and the fitness value is the same as the optimization target value.

9. A scheduling system for precision equipment production workshops considering resource collaboration and equipment reconstruction. Features: The system includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method for scheduling a precision equipment production workshop that considers resource collaboration and equipment reconstruction as described in any one of claims 1 to 8 is executed.

10. A computer-readable storage medium, Features: The computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are called and executed by the processor, the machine-executable instructions prompt the processor to implement the precision equipment production workshop scheduling method considering resource collaboration and equipment reconstruction as described in any one of claims 1-8.

Images