Superspeed optical module production line scheduling method

Abstract

The invention provides an ultra-high-speed optical module production line scheduling method. The method comprises the following steps: S1, establishing a scheduling mathematical model according to a set optimization target and constraint conditions; s2, setting parameters of an ant colony algorithm and reinforcement learning; s3, constructing a state transition rule and heuristic information of the ant colony algorithm, and initializing a pheromone matrix of the ant colony algorithm; s4, initializing the ant colony position and path of the ant colony algorithm; s5, the ant colony algorithm runs according to the pheromone matrix to obtain a scheduling scheme and the maximum completion time of the scheduling scheme; s6, reinforcement learning updates the pheromone matrix based on the maximum completion time, and the ant colony algorithm updates the scheduling scheme according to the updated pheromone matrix; and when the number of iterations of reinforcement learning reaches a set number, outputting a scheduling scheme, otherwise, returning to S4. According to the method, the production efficiency of an ultra-high-speed optical module production line is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ultra-high-speed optical module production line scheduling and intelligent optimization algorithm, and in particular to an ultra-high-speed optical module production line scheduling method. Background Art

[0002] The scheduling problem of ultra-high-speed optical module production lines is a key issue in modern manufacturing, involving multiple workpieces processed by multiple machines in multiple processes. With the advancement of Industry 4.0, the manufacturing industry has entered a new era of intelligence, automation, and informationization, which has put forward higher requirements for production scheduling. In the context of Industry 4.0, how to efficiently schedule production resources and optimize production processes has become a core issue in improving the competitiveness of the manufacturing industry.

[0003] Industry 4.0 advocates the use of digital technology and intelligent means to achieve high automation and flexibility in the production process, thereby improving production efficiency and product quality, shortening production cycles, and quickly responding to market demand. In this context, traditional scheduling methods are stretched to the limit and cannot cope with dynamically changing production environments and complex production requirements. The scheduling problem of ultra-high-speed optical module production lines becomes more complicated in the environment of Industry 4.0, especially in production systems with batch processing characteristics. In many actual production environments, the batch processing capability of the machine has a significant impact on the overall production efficiency. On the one hand, the batch processing feature means that the machine can process multiple workpieces at the same time. This processing method can significantly improve production efficiency, reduce energy consumption and equipment idle time in some scenarios; on the other hand, each machine can process a limited number of workpieces at the same time in one processing cycle, and each workpiece needs to be processed in a specific process sequence. The processing time of different workpieces in different processes may be different, which requires the scheduling strategy to reasonably arrange the batches of workpieces to fully utilize the processing capacity of the machine and flexibly respond to complex processing paths and time arrangements.

[0004] Traditional scheduling methods often face problems such as high computational complexity and difficulty in dynamic adjustment when dealing with the scheduling problem of ultra-high-speed optical module production lines with batch processing characteristics. With the development of manufacturing industry towards intelligence and automation, it has become a trend to use intelligent optimization algorithms to solve scheduling problems. Intelligent optimization algorithms can provide high-quality scheduling solutions in a short time and adapt to dynamically changing production environments. Among intelligent optimization algorithms, the ant colony algorithm, as a heuristic algorithm based on swarm intelligence, has strong global search capabilities and adaptability, and has been widely used in various combinatorial optimization problems. However, when facing complex scheduling problems, the traditional ant colony algorithm still has shortcomings such as slow convergence speed and easy to fall into local optimality. Summary of the invention

[0005] The purpose of the present invention is to provide an ultra-high-speed optical module production line scheduling method to improve scheduling optimization efficiency and enhance production efficiency.

[0006] In order to solve the above technical problems, the present invention provides an ultra-high-speed optical module production line scheduling method, comprising the following steps: S1. Establish a scheduling mathematical model based on the set optimization objectives and constraints; S2, setting the parameters of the ant colony algorithm and reinforcement learning, where the ant colony algorithm is constrained by the scheduling mathematical model; S3, construct the state transition rules and heuristic information of the ant colony algorithm, and initialize the pheromone matrix of the ant colony algorithm; S4, initializing the ant colony position and path of the ant colony algorithm; S5, the ant colony algorithm runs according to the pheromone matrix to obtain the scheduling plan and the maximum completion time of the scheduling plan; S6. Reinforcement learning updates the pheromone matrix based on the maximum completion time, and the ant colony algorithm updates the scheduling plan based on the updated pheromone matrix; when the number of iterations of reinforcement learning reaches the set number, the scheduling plan is output, otherwise it returns to S4. According to the above scheme, the optimization goal is: to minimize the maximum completion time of the scheduling scheme; The constraints include: 1) A certain process of the workpiece can only be started after the previous process is completed; 2) Any process cannot be interrupted after it starts processing; 3) Any process can only be processed once on the machine; 4) The completion time of any process shall not exceed the maximum completion time; 5) A general processing procedure in which a discrete machine processes at most one workpiece at a time, and a batch processing procedure in which a batch processing machine processes at most B workpieces at a time, where B is the maximum number of workpieces that the batch processing machine can process at one time; the processing time of a batch processing procedure on a batch processing machine is the sum of the processing times of the batch processing procedures processed by the batch processing machine; 6) The processing time of a batch of batch processes on a batch processing machine is the sum of the processing time of the batch processes of the batch on the batch processing machine; 7) The start processing time, processing time and completion time of any process are non-negative; 8) The workpieces processed in each batch of the batch processing machine are determined by the order of completion of the previous process of the batch processing process. When the number of completed processes of the previous process of the batch processing process accumulates to the maximum batch processing capacity, it is counted as one batch for processing; if the number of remaining workpieces is less than the maximum batch processing capacity, these remaining workpieces constitute the last batch for processing; 9) The processing machine of the batch process of each batch is the same as the processing machine of the workpiece of the previous process corresponding to the first completed batch process number of the batch.

[0007] According to the above scheme, the pheromone matrix includes a pheromone matrix and a machine pheromone concentration matrix; initializing the pheromone matrix of the ant colony algorithm includes: filling the pheromone matrix and the machine pheromone concentration matrix with a set initial pheromone concentration.

[0008] According to the above scheme, step S4 includes: initializing the initial solution of each ant and setting the initial process sequence according to the characteristics of the scheduling mathematical model.

[0009] According to the above scheme, the scheduling scheme is represented as a Gantt chart, and the maximum completion time is obtained by calculating the path quality of each ant.

[0010] According to the above scheme, step S6 includes: reinforcement learning based on the update amount of elements in the maximum completion time pheromone matrix and the machine pheromone concentration matrix, as well as the pheromone volatilization rate, and then updating the elements in the pheromone matrix and the machine pheromone concentration matrix.

[0011] According to the above scheme, reinforcement learning is a Q-learning algorithm.

[0012] The present invention also provides an ultra-high-speed optical module production line scheduling system, which executes the ultra-high-speed optical module production line scheduling method described above.

[0013] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the ultra-high-speed optical module production line scheduling method described above are implemented.

[0014] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the ultra-high-speed optical module production line scheduling method described above.

[0015] Beneficial Effects Aiming at the characteristics of the scheduling problem of the ultra-high-speed optical module production line, the present invention establishes a scheduling mathematical model, which provides a basis for the subsequent optimization of scheduling schemes; by combining reinforcement learning with the ant colony algorithm, the existing ant colony algorithm is optimized, the search and utilization capabilities of the ant colony algorithm are enhanced, and the optimization efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a method for scheduling an ultra-high-speed optical module production line according to an embodiment of the present invention; Figure 2 is an update flow chart of the ant colony algorithm and reinforcement learning according to an embodiment of the present invention; Figure 3 It is a schematic diagram of process coding of an embodiment of the present invention; Figure 4 is a Gantt chart output by an ant colony algorithm according to an embodiment of the present invention; Figure 5 It is a curve diagram showing the relationship between the iteration rounds and the optimization time of the ant colony algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0018] See also Figure 1 , Figure 2 This embodiment provides a method for scheduling an ultra-high-speed optical module production line, including the following steps: S1. Establish a scheduling mathematical model based on the set optimization objectives and constraints; S2, setting the parameters of the ant colony algorithm and reinforcement learning, where the ant colony algorithm is constrained by the scheduling mathematical model; S3, construct the state transition rules and heuristic information of the ant colony algorithm, and initialize the pheromone matrix of the ant colony algorithm; S4, initializing the ant colony position and path of the ant colony algorithm; S5, the ant colony algorithm runs according to the pheromone matrix to obtain the scheduling plan and the maximum completion time of the scheduling plan; S6. Reinforcement learning updates the pheromone matrix based on the maximum completion time, and the ant colony algorithm updates the scheduling plan based on the updated pheromone matrix; when the number of iterations of reinforcement learning reaches the set number, the scheduling plan is output, otherwise it returns to S4.

[0019] In this embodiment, the ultra-high-speed optical module production line scheduling method is applicable to a hybrid assembly line workshop with batch processing characteristics, wherein the workshop includes batch processing machines and discrete machines, which are respectively used for batch processing procedures and general procedures of processing workpieces.

[0020] In this embodiment, the parameters of the scheduling mathematical model are set as follows.

[0021] : Total number of workpieces; : Workpiece The number of processes; : Total number of machines; : artifact collection; : Workpiece No. A process; : machine collection; : Batch machine collection; : The maximum number of workpieces that a batch processor can process at one time; : Workpiece No. The process in the machine Processing time on : Workpiece No. Start time of the first process: : Workpiece No. The completion time of a process; : Workpiece completion time; : Workpiece No. Is the process on the machine? Processing, for 0-1 variables, in The upper processing is 1; ; Number of batch processing steps; : Batch process number collection; :The first The first process is a batch process; :The first The process is a batch process; ...; :The first The first process is a batch process; : The number of completed workpieces of the previous process corresponding to the batch process number; :The first The number of completed workpieces for a process is ; :The first The number of completed workpieces for a process is ;……; :The first The number of completed workpieces for a process is ; : The matrix of the completed workpiece number of the previous process corresponding to the batch process number of the current processing batch; the matrix elements : No. The current processing batch of the process The number of the completed workpiece.

[0022] Furthermore, the optimization goal is to minimize the maximum completion time of the scheduling scheme;

[0023] The constraints include: 1) Workpiece (workpiece ) of a process ( The first process) must be in the previous process (the It can only start after the first process is completed;

[0024] 2) Any process cannot be interrupted after it starts processing;

[0025] 3) Any process can only be processed once on the machine;

[0026] 4) The completion time of any process shall not exceed the maximum completion time;

[0027] 5) A general processing procedure in which a discrete machine processes at most one workpiece at a time, and a batch processing procedure in which a batch processing machine processes at most B workpieces at a time, where B is the maximum number of workpieces that the batch processing machine can process at one time; the processing time of a batch processing procedure on a batch processing machine is the sum of the processing times of the batch processing procedures processed by the batch processing machine;

[0028]

[0029] 6) The processing time of a batch of batch processes on a batch processing machine is the sum of the processing time of the batch processes of the batch on the batch processing machine;

[0030] 7) The start processing time, processing time and completion time of any process are non-negative;

[0031] 8) The workpieces processed in each batch of the batch processing machine are determined by the order of completion of the previous process of the batch processing process. When the number of completed processes of the previous process of the batch processing process accumulates to the maximum batch processing capacity, it is counted as one batch for processing; if the number of remaining workpieces is less than the maximum batch processing capacity, these remaining workpieces constitute the last batch for processing;

[0032] Record the completed workpiece number of the previous process corresponding to the batch process number of the current processing batch in the order of completion; when the number of completed workpieces of the previous process corresponding to the batch process number reaches an integer multiple of the maximum batch processing capacity or the total number of workpieces, clear the record; 9) The processing machine of each batch of batch processing procedures is the same as the processing machine of the workpiece of the previous process corresponding to the first completed batch processing procedure number of the batch; .

[0033] Furthermore, the parameters of the ant colony algorithm include: the number of ants am, the maximum number of iterations t, the pheromone weight factor , heuristic information weight factor , pheromone volatility coefficient ; The parameters of reinforcement learning include: learning rate , Exploration Rate , Discount Factor .

[0034] Furthermore, the state transition rule includes the ant from a certain process (process ) is transferred to another process (process ) state transition probability , the state transition probability is determined by the pheromone concentration and heuristic information get;

[0035] Heuristic information through pheromone volatility coefficient , pheromone deposition get;

[0036] Pheromone deposition according to ant The quality of the found path is obtained; ;

[0037] In the above formula, It's an ant The total length of the path traversed, is the pheromone deposition constant.

[0038] Furthermore, the pheromone matrix includes a pheromone matrix and machine pheromone concentration matrix ; Initializing the pheromone matrix of the ant colony algorithm includes: filling the pheromone matrix with the set initial pheromone concentration and machine pheromone concentration matrix ; Pheromone Matrix It is expressed as:

[0039] The number of rows is the number of workpiece processes , the number of columns is the product of the number of workpieces and the number of processes A matrix whose elements Indicates that the ants are in topological order. Process to Select when a process pheromone concentration as a path; By machine pheromone concentration matrix is divided into m sub-matrices, and the rows and columns of each sub-matrix are and , each sub-matrix is ​​expressed as:

[0040]

[0041]

[0042]

[0043] The number of rows is the total number of artifacts , the number of columns is the number of workpiece processes , the depth is the total number of machines A matrix whose elements Representation of workpiece Process In the machine The concentration of processed pheromones on.

[0044] Furthermore, step S4 includes: initializing the initial solution of each ant, and setting the initial process sequence according to the characteristics of the scheduling mathematical model.

[0045] In this embodiment, setting the initial process sequence includes allocating batch processing machine processes and discrete machine processes. The initial process sequence is represented by a single-layer coding rule based on the workpiece processing sequence; see Figure 3The numbers in the single-layer coding rule are the workpiece numbers. The number of times the numbers are repeated indicates the corresponding process of the workpiece. The processing order follows the coding order, for example: [1 3 2 33 2 1 2 2 3 1 1], where the first 3 represents the first process of the workpiece with serial number 3, and the second 3 represents the second process of the workpiece with serial number 3.

[0046] Further, see Figure 4 ,The scheduling scheme is represented as a Gantt chart, and the maximum completion time is obtained by calculating the path quality of each ant.

[0047] Further, step S6 includes: reinforcement learning based on the maximum completion time pheromone matrix and machine pheromone concentration matrix Update the amount of elements in the pheromone matrix and the pheromone volatilization rate and machine pheromone concentration matrix The elements in; its update process is expressed as: ;

[0048] In the above formula, , Pheromone Matrix and machine pheromone concentration matrix The update amount of the elements in , , Pheromone Matrix The values ​​of the elements before and after the update, , The machine pheromone concentration matrix is The values ​​of the elements before and after the update, is the pheromone volatilization rate obtained by reinforcement learning.

[0049] Furthermore, reinforcement learning is a Q-learning algorithm; the update formula of Q-learning is:

[0050] In the above formula, and Represent the current state and the next state respectively. and Represents the current action and the next action respectively. is the reward value; the update effect of reinforcement learning on the ant colony algorithm in this embodiment can be found in Figure 5 .

[0051] The ultra-high-speed optical module production line scheduling method of this embodiment has at least the following advantages: (1) According to the scheduling characteristics of the ultra-high-speed optical module production line scheduling problem of a batch processing machine with the characteristic of maximizing batch processing capacity, a mathematical model suitable for the ultra-high-speed optical module production line scheduling problem of a batch processing machine with the characteristic of maximizing batch processing capacity is proposed, and the processing sequence and time constraints of the workpieces are defined.

[0052] (2) The state transition rules and heuristic information of the ant colony algorithm were designed, and the Q-learning mechanism was introduced to dynamically adjust the pheromone volatilization rate and update rate, thereby enhancing the algorithm's search and utilization capabilities.

[0053] (3) The path pheromone concentration matrix and machine pheromone concentration matrix were constructed to represent the pheromone concentrations between processes and on machines, respectively.

[0054] (4) Simulation experiments are used to verify the effectiveness and robustness of the proposed method on scheduling problems of different scales.

[0055] (5) The algorithm improves scheduling efficiency through intelligent optimization and makes full use of the batch processing capacity of the machine. It is suitable for the specific complex production environment of the ultra-high-speed optical module production line scheduling problem of batch processing machines with the characteristics of maximizing batch processing capacity. It significantly reduces the maximum completion time and improves production efficiency. It provides an efficient and feasible solution to the ultra-high-speed optical module production line scheduling problem of batch processing machines with the characteristics of maximizing batch processing capacity, and has broad application prospects.

[0056] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, and two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0057] 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 method for scheduling an ultra-high-speed optical module production line, characterized in that: The following steps are involved: S1. Establish a scheduling mathematical model based on the set optimization objectives and constraints; S2, setting the parameters of the ant colony algorithm and reinforcement learning, where the ant colony algorithm is constrained by the scheduling mathematical model; S3, construct the state transition rules and heuristic information of the ant colony algorithm, and initialize the pheromone matrix of the ant colony algorithm; S4, initializing the ant colony position and path of the ant colony algorithm; S5, the ant colony algorithm runs according to the pheromone matrix to obtain the scheduling plan and the maximum completion time of the scheduling plan; S6. Reinforcement learning updates the pheromone matrix based on the maximum completion time, and the ant colony algorithm updates the scheduling plan based on the updated pheromone matrix; when the number of iterations of reinforcement learning reaches the set number, the scheduling plan is output, otherwise it returns to S4.

2. The ultra-high-speed optical module production line scheduling method according to claim 1, characterized in that: The optimization goal is to minimize the maximum completion time of the scheduling scheme; The constraints include: 1) A certain process of the workpiece can only be started after the previous process is completed; 2) Any process cannot be interrupted after it starts processing; 3) Any process can only be processed once on the machine; 4) The completion time of any process shall not exceed the maximum completion time; 5) A general processing procedure in which a discrete machine processes at most one workpiece at a time, and a batch processing procedure in which a batch processing machine processes at most B workpieces at a time, where B is the maximum number of workpieces that the batch processing machine can process at one time; the processing time of a batch processing procedure on a batch processing machine is the sum of the processing times of the batch processing procedures processed by the batch processing machine; 6) The processing time of a batch of batch processes on a batch processing machine is the sum of the processing time of the batch processes of the batch on the batch processing machine; 7) The start processing time, processing time and completion time of any process are non-negative; 8) The workpieces processed in each batch of the batch processing machine are determined by the order of completion of the previous process of the batch processing process. When the number of completed processes of the previous process of the batch processing process accumulates to the maximum batch processing capacity, it is counted as one batch for processing; if the number of remaining workpieces is less than the maximum batch processing capacity, these remaining workpieces constitute the last batch for processing; 9) The processing machine of the batch process of each batch is the same as the processing machine of the workpiece of the previous process corresponding to the first completed batch process number of the batch.

3. The ultra-high-speed optical module production line scheduling method according to claim 1, characterized in that: The pheromone matrix includes a pheromone matrix and a machine pheromone concentration matrix; initializing the pheromone matrix of the ant colony algorithm includes: filling the pheromone matrix and the machine pheromone concentration matrix with a set initial pheromone concentration.

4. The ultra-high-speed optical module production line scheduling method according to claim 1, characterized in that: Step S4 includes: initializing the initial solution of each ant and setting the initial process sequence according to the characteristics of the scheduling mathematical model.

5. The ultra-high-speed optical module production line scheduling method according to claim 1, characterized in that: The scheduling scheme is represented as a Gantt chart, and the maximum completion time is obtained by calculating the path quality of each ant.

6. The ultra-high-speed optical module production line scheduling method according to claim 3, characterized in that: Step S6 includes: strengthening learning based on the update amount of elements in the maximum completion time pheromone matrix and the machine pheromone concentration matrix, and the pheromone volatilization rate, and then updating the elements in the pheromone matrix and the machine pheromone concentration matrix.

7. The ultra-high-speed optical module production line scheduling method according to claim 1, characterized in that: Reinforcement learning is the Q-learning algorithm.

8. An ultra-high-speed optical module production line scheduling system, characterized in that: The system executes the ultra-high-speed optical module production line scheduling method described in any one of claims 1-7.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the ultra-high-speed optical module production line scheduling method described in any one of claims 1-7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the ultra-high-speed optical module production line scheduling method described in any one of claims 1-7 are implemented.

Images