Order ranking and processing scheduling method considering machine maintenance
Patent Information
- Application Number
- CN202211388242.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-08
- Publication Date
- 2026-11-03
- Estimated Expiration
- 2042-11-08
AI Technical Summary
Traditional machine maintenance methods are inefficient in production scheduling, resulting in high production management costs and low production efficiency for enterprises, making it difficult to cope with changing product models and complex batch requirements.
An order hierarchical sorting and processing scheduling method is adopted. By classifying orders into urgent, normal and low-profit levels, the reliability and failure rate of machines are calculated using the Weibull distribution, a mathematical model is established, and intelligent optimization algorithms are combined to optimize machine maintenance strategies, including daily inspection, preventive maintenance and fault repair, to generate the optimal scheduling scheme.
Maximize production profits within a limited timeframe, minimize maintenance costs and delay losses, optimize order processing completion times, improve production efficiency, and reduce management costs.
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent manufacturing, and more specifically to a method for order hierarchical sorting and processing scheduling that takes machine maintenance into account. Background Technology
[0002] As domestic enterprises gradually digitize and automate their machines, traditional, reactive machine maintenance methods are leaving production managers struggling to efficiently utilize machines. For example, many standard parts manufacturers currently manage production scheduling in a very rudimentary way. Faced with an increasing number of product models and complex, variable production batches, schedulers are finding it difficult to rationally arrange production schedules. These problems hinder standard parts companies from improving their efficiency. To cope with a more open market environment, optimizing internal management, solving existing production management problems, and improving production efficiency while reducing production management costs have become inevitable choices for Chinese enterprises. Therefore, the importance of integrating and researching equipment maintenance and production scheduling management, which are crucial aspects of production management, is self-evident. Summary of the Invention
[0003] This invention provides a hierarchical ordering and processing scheduling method that takes machine maintenance into account, in order to achieve the optimization goals of maximizing production profits, minimizing maintenance costs, minimizing delay losses, and minimizing order processing completion time within a finite time domain.
[0004] To achieve the above objectives, the technical solution adopted by this invention mainly includes the following processes:
[0005] Step 1: Based on the order's delivery time, production profit, and deferred loss profit, classify the orders into urgent, normal, and low-profit levels. Sort and number the orders in each level. Break down all orders into multiple sub-orders, sort and number the sub-orders, thereby generating the order sequence for each level. Number the optional processing machines for all processes in each sub-order.
[0006] The reliability of each processing machine and the failure rate of its important components are calculated using the Weibull distribution. The linear combination of reliability and failure rate is used as the health value evaluation function for each processing machine. For all processing machines that can be selected for each process, the values of their health value evaluation functions are divided into three intervals from large to small, corresponding to three machine sets: high health set, medium health set, and low health set.
[0007] Among the optional processing machines for the same process, the interval of the health value evaluation function value corresponding to the high-health group is [R1, R2]. m The interval of the health value evaluation function value corresponding to the moderate health group is [R]. m+1 R nThe interval of the health value evaluation function value corresponding to the low health group is [R]. n+1 R K ], where R1>R m >R m+1 >R n >R n+1 >R K ;
[0008] Step 2: Based on the known order and machine information, establish a mathematical model with the objective functions of maximizing production profit, minimizing maintenance costs, minimizing delay losses, and minimizing order processing completion time within a finite time domain as optimization objectives;
[0009] The model is as follows: The workshop has multiple batches of machines with different performance (M1, M2, M3, ..., M...). k M K ), loading station (M) O ), unloading station (M) K+1 It consists of a parts library and an order sorting sequence at each level {L}. 11 L 12 , ..., L 1l L 21 , ..., L hl The type and quantity of the order are known (L) hln —Quantity of the l-th sub-order in order number h), the order's process is a hlij The cost of each order is known (c) hl - The cost corresponding to the l-th sub-order in order number h); the machine maintenance types in the overall production stage are routine inspection, preventive maintenance, and fault repair, and their single maintenance costs are C respectively. P1 C P2 C P3 The time for preventive maintenance and fault repair is t. p2 t p3 The processing time for the process is t. khlij (Machine M) k The processing time required for order number hl); the profit per order is P. hlc The loss profit for a single order is P. hlx ;
[0010] Step 3: Treat all orders of the same level as a whole, set scheduling nodes according to the maximum completion time and order sequence, and treat the process between two adjacent scheduling nodes as a scheduling sub-scheme;
[0011] Perform the above operations on all orders of the emergency level, normal level, and low-profit level respectively to generate three different scheduling sub-schemes for orders of different levels, which correspond to the scheduling sub-schemes of type I, type II, and type III respectively.
[0012] Step 4: Set the following rules for the three maintenance types in machine maintenance: routine inspection, pre-maintenance, and fault repair:
[0013] When performing routine maintenance on a machine, the value of its health evaluation function remains unchanged, and no maintenance time needs to be inserted; when performing pre-maintenance on a machine, the value of its health evaluation function will change, and corresponding maintenance time needs to be inserted; when performing fault repair on a machine, the value of its health evaluation function will change, and the status of the machine in this production cycle will be set to unselectable.
[0014] When a processing machine is handling an urgent order, only routine maintenance needs to be considered during the current production cycle. When a processing machine is handling a normal order, both routine maintenance and pre-maintenance need to be considered during the production cycle. When a processing machine is handling a low-profit order, all three maintenance operations—routine maintenance, pre-maintenance, and fault repair—need to be considered during the production cycle.
[0015] Define a fault repair threshold D P3 If the number of maintenance operations for a machine exceeds a certain threshold, it needs to enter a fault repair state. The formation process of the fault repair is as follows: when the machine is pre-maintained, there is a life reduction factor, which means that the machine will not reach its initial service life after the pre-maintainment is performed. As the number of pre-maintainment operations increases, the effective service life of the machine after the pre-maintainment will gradually decrease. When a certain state is reached, pre-maintaining can no longer meet the requirements for putting the machine into production. Therefore, this state is set as fault repair. By analyzing the data of the machine's previous maintenance, the fault repair threshold is finally defined.
[0016] Define a failure rate threshold D K If the failure rate of critical machine components exceeds this threshold, a preparation time t is inserted after the processing step. pij ;
[0017] Define a pre-maintenance threshold D p The value of this threshold is the lower bound of the health value evaluation function interval corresponding to the low-health group, i.e., D. p =R K When the health value evaluation function of a machine is lower than the pre-maintenance threshold, pre-maintenance time is inserted before the processing step.
[0018] Step 5: Use multi-level combined integer encoding to encode the sub-order machine selection sequence, sub-order processing sequence, maintenance time and node insertion sequence, preparation time and node insertion sequence, and total order processing sequence in the same level of order sequence.
[0019] The maintenance time and node insertion sequence are used to represent the maintenance status of the machine after completing the current process. Each position in the sequence corresponds to a process number, and its maintenance status is represented by three different labels: if the machine continues processing after completing the current process, it is represented by label 1; if the machine needs to enter pre-maintenance after completing the current process, it is represented by label 2; if the machine enters fault repair after completing the current process, it is represented by label 3. After the process is completed, the corresponding pre-maintenance time or fault repair time is inserted according to the different labels.
[0020] The maintenance time and the method for constructing the node insertion sequence are as follows:
[0021] Initialize a sequence of the same length as the sub-order processing sequence, where each position in the sequence corresponds to a sub-order.
[0022] The process number in the order processing sequence is used to determine the value of the health value evaluation function of the machine when processing the process set;
[0023] If the value is greater than the pre-maintenance threshold, the machine continues to process, and the corresponding position of the sequence is marked with label 1 until the value of the machine health value evaluation function is less than or equal to the pre-maintenance threshold after a certain process is completed. Then, the machine status when the process starts is set to non-selectable. When the pre-maintenance is completed, the machine status is set to selectable, and the corresponding position of the sequence is marked with label 2.
[0024] If a machine enters fault repair after completing a certain process, and the machine's pre-maintenance count reaches a threshold, the machine's state when the previous process was completed is set to an unselectable state. Then, label 3 is marked at the corresponding position in the sequence, thus obtaining the maintenance time and node insertion sequence.
[0025] The preparation time and node insertion sequence are used to determine whether a sudden failure occurs after the machine completes the current process. Each position in the sequence corresponds to the process number, and the determination method is represented by two different labels: if the machine continues processing after completing the current process, it is represented by label 1a; if the machine experiences a sudden failure after completing the current process and needs to insert preparation time, it is represented by label 1b. After the process is completed, the corresponding preparation time is inserted according to the label.
[0026] The preparation time and the method for constructing the node insertion sequence are as follows:
[0027] Initialize a sequence with the same length as the sub-order processing sequence, where each position in the sequence corresponds to the process number of the sub-order processing sequence, and determine the failure rate of the machine's critical components;
[0028] If the failure rate of a critical component of the machine is lower than the failure rate threshold, the machine continues processing, and the corresponding position in the sequence is marked with label 1a. This continues until a certain process is completed and the failure rate of the critical component of the machine is higher than or equal to the failure rate threshold. Then, the following judgment and operation are performed: when the machine's idle time is greater than or equal to the preparation time, label 1a is marked in the corresponding position in the sequence; when the machine's idle time is less than the preparation time, preparation time needs to be inserted before workpiece processing, and label 1b is marked in the corresponding position in the sequence, thus obtaining the preparation time and node insertion sequence.
[0029] Step 6: Utilize intelligent optimization algorithms to optimize and solve the scheduling sub-schemes for types I, II, and III respectively. The steps include:
[0030] Step 6a: Encode the first type of scheduling sub-scheme using the encoding method of step 5. Optimize and solve the first type of scheduling sub-scheme using an intelligent optimization algorithm. After the urgent orders are processed, update the health value evaluation function of the high health group, classify the machines in the high health group into the medium health group, and update the medium health group.
[0031] Step 6b: Encode the second type of scheduling sub-scheme using the encoding method of step 5. Optimize the second type of scheduling sub-scheme using the intelligent optimization algorithm. After the normal level orders are processed, update the health value evaluation function of the medium health group, reassign the machines in the medium health group to the low health group, and update the low health group.
[0032] Step 6c: Encode the third type of scheduling sub-scheme using the encoding method of step 5, and use the intelligent optimization algorithm to optimize and solve the third type of scheduling sub-scheme. After the low-profit level orders are processed, update the value of the health value evaluation function of the low health group.
[0033] Step 7: Combine the scheduling sub-schemes of type I, type II and type III to complete the scheduling of all orders and generate the overall scheduling scheme to achieve the optimization goals of maximizing production profit, minimizing maintenance costs, minimizing delay losses and minimizing order processing completion time within a finite time domain.
[0034] The assumptions of this invention are described as follows:
[0035] (1) A process that can only process one workpiece at a time on the same machine; a process that can only be processed on one machine at a time.
[0036] (2) There is no priority relationship between the optional process routes and optional machines for the workpiece;
[0037] (3) The occurrence of machine failures follows a Weibull distribution. As time increases, the service life of the machine decreases, the failure rate increases, the health value decreases, and the frequency of preventive maintenance increases.
[0038] (4) Once a process begins, it cannot be interrupted before it is completed. Pre-maintenance of the machine is carried out after a certain process is completed.
[0039] (5) All machines shall remain powered on until all work is completed.
[0040] The objective function of the production scheduling model is as follows:
[0041] Production profit: In the formula, C represents the maintenance cost, and L... bln L represents the number of overdue orders. h-bln The quantity of orders to be completed;
[0042] Completion time: f2 = ∑a hlij +n k t p2 +n g t p3 +∑t qij +∑t pij In the formula ∑a hlij ∑t represents the processing time for this order. qij ∑t represents the idle time allowed for the machine during the processing of this order. pij n represents the machine preparation time during the processing of this order. k n g These represent the number of times the machine underwent pre-maintenance and fault repair during the processing of this order;
[0043] Maintenance costs: In the formula, n1, n2, and n3 represent the number of routine inspections, pre-maintenance, and fault repairs, respectively.
[0044] Delay results in lost profits: Detailed Implementation
[0045] The NSGA-II algorithm is used below to optimize the sub-scheduling scheme, and the specific implementation of the present invention is described in detail below:
[0046] Step 1: Based on the order's delivery time, production profit, and deferred loss profit, classify the orders into urgent, normal, and low-profit levels. Sort and number the orders in each level. Break down all orders into multiple sub-orders, sort and number the sub-orders, thereby generating the order sequence for each level. Number the optional processing machines for all processes in each sub-order.
[0047] The reliability of each processing machine and the failure rate of its important components are calculated using the Weibull distribution. The linear combination of reliability and failure rate is used as the health value evaluation function for each processing machine. For all processing machines that can be selected for each process, the values of their health value evaluation functions are divided into three intervals from large to small, corresponding to three machine sets: high health set, medium health set, and low health set.
[0048] Among the optional processing machines for the same process, the interval of the health value evaluation function value corresponding to the high-health group is [R1, R2]. m The interval of the health value evaluation function value corresponding to the moderate health group is [R]. m+1 R n The interval of the health value evaluation function value corresponding to the low health group is [R]. n+1 R K ], where R1>R m >R m+1 >R n >R n+1 >R K ;
[0049] Step 2: Based on the known order and machine information, establish a mathematical model with the objective functions of maximizing production profit, minimizing maintenance costs, minimizing delay losses, and minimizing order processing completion time within a finite time domain as optimization objectives;
[0050] The model is as follows: The workshop has multiple batches of machines with different performance (M1, M2, M3, ..., M...). k M K ), loading station (M) O ), unloading station (M) K+1 It consists of a parts library and an order sorting sequence at each level {L}. 11 L 12 , ..., L 1l L 21 , ..., L hl The type and quantity of the order are known (L) hln —Quantity of the l-th sub-order in order number h), the order's process is a hlij The cost of each order is known (c) hl- The cost corresponding to the l-th sub-order in order number h); the machine maintenance types in the overall production stage are routine inspection, preventive maintenance, and fault repair, and their single maintenance costs are C respectively. P1 C P2 C P3 The time for preventive maintenance and fault repair is t. p2 t p3 The processing time for the process is t. khlij (Machine M) k The processing time required for order number hl); the profit per order is P. hlc The loss profit for a single order is P. hlx ;
[0051] Step 3: Treat all orders of the same level as a whole, set scheduling nodes according to the maximum completion time and order sequence, and treat the process between two adjacent scheduling nodes as a scheduling sub-scheme;
[0052] Perform the above operations on all orders of the emergency level, normal level, and low-profit level respectively to generate three different scheduling sub-schemes for orders of different levels, which correspond to the scheduling sub-schemes of type I, type II, and type III respectively.
[0053] Step 4: Set the following rules for the three maintenance types in machine maintenance: routine inspection, pre-maintenance, and fault repair:
[0054] When performing routine maintenance on a machine, the value of its health evaluation function remains unchanged, and no maintenance time needs to be inserted; when performing pre-maintenance on a machine, the value of its health evaluation function will change, and corresponding maintenance time needs to be inserted; when performing fault repair on a machine, the value of its health evaluation function will change, and the status of the machine in this production cycle will be set to unselectable.
[0055] When a processing machine is handling an urgent order, only routine maintenance needs to be considered during the current production cycle. When a processing machine is handling a normal order, both routine maintenance and pre-maintenance need to be considered during the production cycle. When a processing machine is handling a low-profit order, all three maintenance operations—routine maintenance, pre-maintenance, and fault repair—need to be considered during the production cycle.
[0056] Define a fault repair threshold D P3If the number of maintenance operations for a machine exceeds a certain threshold, it needs to enter a fault repair state. The formation process of the fault repair is as follows: when the machine is pre-maintained, there is a life reduction factor, which means that the machine will not reach its initial service life after the pre-maintainment is performed. As the number of pre-maintainment operations increases, the effective service life of the machine after the pre-maintainment will gradually decrease. When a certain state is reached, pre-maintaining can no longer meet the requirements for putting the machine into production. Therefore, this state is set as fault repair. By analyzing the data of the machine's previous maintenance, the fault repair threshold is finally defined.
[0057] Define a failure rate threshold D K If the failure rate of critical machine components exceeds this threshold, a preparation time t is inserted after the processing step. pij ;
[0058] Define a pre-maintenance threshold D p The value of this threshold is the lower bound of the health value evaluation function interval corresponding to the low-health group, i.e., D. p =R K When the health value evaluation function of a machine is lower than the pre-maintenance threshold, pre-maintenance time is inserted before the processing step.
[0059] Step 5: Use multi-level combined integer encoding to encode the sub-order machine selection sequence, sub-order processing sequence, maintenance time and node insertion sequence, preparation time and node insertion sequence, and total order processing sequence in the same level of order sequence.
[0060] The maintenance time and node insertion sequence are used to represent the maintenance status of the machine after completing the current process. Each position in the sequence corresponds to a process number, and its maintenance status is represented by three different labels: if the machine continues processing after completing the current process, it is represented by label 1. This label is only used for identification and can be changed to other numbers; if the machine needs to enter pre-maintenance after completing the current process, it is represented by label 2; if the machine enters fault repair after completing the current process, it is represented by label 3. After the process is completed, the corresponding pre-maintenance time or fault repair time is inserted according to the different labels.
[0061] The maintenance time and the method for constructing the node insertion sequence are as follows:
[0062] Initialize a sequence of the same length as the sub-order processing sequence, where each position in the sequence corresponds to a sub-order.
[0063] The process number in the order processing sequence is used to determine the value of the health value evaluation function of the machine when processing the process set;
[0064] If the value is greater than the pre-maintenance threshold, the machine continues to process and marks the corresponding position of the sequence with label 1. This label is only used for identification and can be changed to other numbers. When the value of the machine health value evaluation function is less than or equal to the pre-maintenance threshold after a certain process is completed, the machine status when the process started is set to non-selectable. When the pre-maintenance is completed, the machine status is set to selectable and the corresponding position of the sequence is marked with label 2.
[0065] If a machine enters fault repair after completing a certain process, and the machine's pre-maintenance count reaches a threshold, the machine's state when the previous process was completed is set to an unselectable state. Then, label 3 is marked at the corresponding position in the sequence, thus obtaining the maintenance time and node insertion sequence.
[0066] The preparation time and node insertion sequence are used to determine whether a sudden failure occurs after the machine completes the current process. Each position in the sequence corresponds to the process number, and the determination method is represented by two different labels: if the machine continues processing after completing the current process, it is represented by label 1a; if the machine experiences a sudden failure after completing the current process and needs to insert preparation time, it is represented by label 1b. After the process is completed, the corresponding preparation time is inserted according to the label.
[0067] The preparation time and the method for constructing the node insertion sequence are as follows:
[0068] Initialize a sequence with the same length as the sub-order processing sequence, where each position in the sequence corresponds to the process number of the sub-order processing sequence, and determine the failure rate of the machine's critical components;
[0069] If the failure rate of a critical component of the machine is lower than the failure rate threshold, the machine continues processing, and the corresponding position in the sequence is marked with label 1a. This continues until a certain process is completed and the failure rate of the critical component of the machine is higher than or equal to the failure rate threshold. Then, the following judgment and operation are performed: when the machine's idle time is greater than or equal to the preparation time, label 1a is marked in the corresponding position in the sequence; when the machine's idle time is less than the preparation time, preparation time needs to be inserted before workpiece processing, and label 1b is marked in the corresponding position in the sequence, thus obtaining the preparation time and node insertion sequence.
[0070] Step 6: Utilize intelligent optimization algorithms to optimize and solve the scheduling sub-schemes for types I, II, and III respectively. The steps include:
[0071] Step 6a: Encode the first type of scheduling sub-scheme using the encoding method of step 5. Optimize and solve the first type of scheduling sub-scheme using the NSGA-II algorithm. After the urgent orders are processed, update the health value evaluation function of the high health group, classify the machines in the high health group into the medium health group, and update the medium health group.
[0072] Step 6b: Encode the Type II scheduling sub-scheme using the encoding method of Step 5. Optimize and solve the Type II scheduling sub-scheme using the NSGA-II algorithm. After the normal level orders are processed, update the health value evaluation function of the medium health group, divide the machines in the medium health group into the low health group, and update the low health group.
[0073] Step 6c: Encode the third type of scheduling sub-scheme using the encoding method of step 5, and use the NSGA-II algorithm to optimize and solve the third type of scheduling sub-scheme. After the low-profit level orders are processed, update the value of the health value evaluation function of the low health group.
[0074] Step 7: Combine the scheduling sub-schemes of type I, type II and type III to complete the scheduling of all orders and generate the overall scheduling scheme to achieve the optimization goals of maximizing production profit, minimizing maintenance costs, minimizing delay losses and minimizing order processing completion time within a finite time domain.
[0075] The objective function of the production scheduling model is as follows:
[0076] Production profit: In the formula, C represents the maintenance cost, and L... bln L represents the number of overdue orders. h-bln The quantity of orders to be completed;
[0077] Completion time: f2 = ∑a hlij +n k t p2 +n g t p3 +∑t qij +∑t pij In the formula ∑a hlij ∑t represents the processing time for this order. qij ∑t represents the idle time allowed for the machine during the processing of this order. pij n represents the machine preparation time during the processing of this order. k n g These represent the number of times the machine underwent pre-maintenance and fault repair during the processing of this order;
[0078] Maintenance costs: In the formula, n1, n2, and n3 represent the number of routine inspections, pre-maintenance, and fault repairs, respectively.
[0079] Delay results in lost profits:
[0080] The above-mentioned encoding method can be solved using various existing mainstream intelligent optimization algorithms. The NSGA-II algorithm is used as an example below.
[0081] The crowding distance between the first and last individuals in each rank solution is set to infinity. The crowding distance of each individual in the same rank solution is the sum of the distance differences between its two adjacent individuals on each sub-objective function, calculated using the following formula:
[0082] D j =(f j+1，1 -f j-1，1 )+(f j+1，2 -f j-1，2 )+(f j+1，3 -f j-1，3 )+(f j+1，4 -f j-1，4 )
[0083] The specific process of the NSGA-II algorithm is as follows:
[0084] S1. Initialize the parameters of the NSGA-II algorithm: population size N, number of iterations G, crossover probability P c Probability of mutation P m And set the current algebra i = 0;
[0085] S2, Randomly generate the initial population P i And complete the fast non-dominated sort;
[0086] S3, regarding the initial population P i Use crossover and mutation operations to generate a offspring population Q. i ;
[0087] S4, P i and Q i Merge, i.e., V i =P i ∪Q i Based on Pareto's dominance and unrelated relationships, a fast non-dominated ranking method is used to construct V. i A set of non-dominated solutions of different levels, Y1, Y2, Y3, ...;
[0088] S5, Clear P i+1 ;
[0089] S6. Calculate the crowding degree of the graded non-dominated solution set, sort according to the non-dominated solution set rank and the crowding distance, and add the first N individuals to P i+1 ;
[0090] S7. Perform crossover and mutation operations on P i+1 to generate the offspring population Q i+1 ;
[0091] S8. If i<G, then i=i+1, return to step S4; otherwise, the algorithm terminates.
[0092] S9. Use the AHP analytic hierarchy process to select a satisfactory solution from the final Pareto optimal solutions.
Claims
1. A method for hierarchical ordering and processing scheduling considering machine maintenance, characterized in that: 1) An order hierarchical sorting and processing scheduling method that takes machine maintenance into account is used for production scheduling in a processing workshop. The workshop is equipped with multiple batches of processing machines with different performance characteristics. The workshop includes multiple parts preparation warehouses, repair stations, loading stations and unloading stations. 2) A method for hierarchical ordering and processing scheduling that considers machine maintenance, comprising the following steps: Step 1: Based on the order's delivery time, production profit, and deferred loss profit, classify the orders into urgent, normal, and low-profit levels. Sort and number the orders in each level. Break down all orders into multiple sub-orders, sort and number the sub-orders, thereby generating the order sequence for each level. Number the optional processing machines for all processes in each sub-order. The reliability of each processing machine and the failure rate of its important components are calculated using the Weibull distribution. The linear combination of reliability and failure rate is used as the health value evaluation function for each processing machine. For all processing machines that can be selected for each process, the values of their health value evaluation functions are divided into three intervals from large to small, corresponding to three machine sets: high health set, medium health set, and low health set. Step 2: Based on the known order and machine information, establish a mathematical model with the objective functions of maximizing production profit, minimizing maintenance costs, minimizing delay losses, and minimizing order processing completion time within a finite time domain as optimization objectives; Step 3: Perform the following operations on all orders of the urgency, normal, and low-profit levels respectively: Treat all orders of the same level as a whole, set scheduling nodes according to the maximum completion time and order sequence, take the process between two adjacent scheduling nodes as a scheduling sub-scheme, and generate scheduling sub-schemes for the three different levels of orders of urgency, normal, and low-profit levels, which correspond to the first, second, and third types of scheduling sub-schemes respectively. Step 4: Set the following rules for the three maintenance types in machine maintenance: routine inspection, pre-maintenance, and fault repair: When performing routine maintenance on a machine, the value of its health evaluation function remains unchanged, and no maintenance time needs to be inserted; when performing pre-maintenance on a machine, the value of its health evaluation function will change, and corresponding maintenance time needs to be inserted; when performing fault repair on a machine, the value of its health evaluation function will change, and the status of the machine in this production cycle will be set to unselectable. When a processing machine is handling an urgent order, only routine maintenance needs to be considered during the current production cycle. When a processing machine is handling a normal order, both routine maintenance and pre-maintenance need to be considered during the production cycle. When a processing machine is handling a low-profit order, all three maintenance operations—routine maintenance, pre-maintenance, and fault repair—need to be considered during the production cycle. Define a fault repair threshold. If the number of maintenance times for a machine exceeds this threshold, it needs to enter a fault repair state. Define a failure rate threshold. If the failure rate of a critical component of the machine exceeds this threshold, insert preparation time after the current process is completed. Define a pre-maintenance threshold, the size of which is the lower bound of the health value evaluation function interval corresponding to the low health set; when the value of the health value evaluation function of a machine is lower than the pre-maintenance threshold, pre-maintenance time is inserted after the current process is completed; Step 5: Using a multi-level combined integer encoding method, encode the sub-order machine selection sequence, sub-order processing sequence, maintenance time and node insertion sequence, preparation time and node insertion sequence, and total order processing sequence in the same level of order sequence. The maintenance time and node insertion sequence are used to represent the maintenance status of the machine after completing the current process. Each position in the sequence corresponds to a process number, and its maintenance status is represented by three different labels: if the machine continues processing after completing the current process, it is represented by label 1; if the machine needs to enter pre-maintenance after completing the current process, it is represented by label 2; if the machine enters fault repair after completing the current process, it is represented by label 3. After the process is completed, the corresponding pre-maintenance time or fault repair time is inserted according to the different labels. The preparation time and node insertion sequence are used to determine whether a sudden failure occurs after the machine completes the current process. Each position in the sequence corresponds to the process number, and the determination method is represented by two different labels: if the machine continues processing after completing the current process, it is represented by label 1a; if the machine experiences a sudden failure after completing the current process and needs to insert preparation time, it is represented by label 1b; after the process is completed, the corresponding preparation time is inserted according to the label. Step 6: Utilize intelligent optimization algorithms to optimize and solve the scheduling sub-schemes for types I, II, and III respectively. The steps include: Step 6a: Encode the first type of scheduling sub-scheme using the encoding method of step 5. Optimize and solve the first type of scheduling sub-scheme using an intelligent optimization algorithm. After the urgent orders are processed, update the health value evaluation function of the high health group, classify the machines in the high health group into the medium health group, and update the medium health group. Step 6b: Encode the second type of scheduling sub-scheme using the encoding method of step 5. Optimize the second type of scheduling sub-scheme using the intelligent optimization algorithm. After the normal level orders are processed, update the health value evaluation function of the medium health group, reassign the machines in the medium health group to the low health group, and update the low health group. Step 6c: Encode the third type of scheduling sub-scheme using the encoding method of step 5, and use the intelligent optimization algorithm to optimize and solve the third type of scheduling sub-scheme. After the low-profit level orders are processed, update the value of the health value evaluation function of the low health group. Step 7: Combine the scheduling sub-schemes of type I, type II and type III to complete the scheduling of all orders and generate the overall scheduling scheme.
2. The order hierarchical sorting and processing scheduling method considering machine maintenance according to claim 1, characterized in that, The objective function for maximizing production profit within the finite time domain, as described in step 2, is constructed in the following way: The objective function for maximizing production profit within a finite time domain is the sum of the evaluation functions for order completion rate and order revenue, production and maintenance costs, and delayed order losses. The order completion rate and order revenue evaluation function is the sum of the profits of all completed orders within a finite time period; the production and maintenance cost evaluation function consists of a production cost evaluation function and a maintenance cost evaluation function: the production cost evaluation function is the product of the production cost of a single component and the number of components, and the maintenance cost evaluation function is the product of the maintenance cost and the number of maintenance operations, with the maintenance cost depending on the maintenance type; the delayed order loss evaluation function is the sum of the delayed loss profits of the total number of uncompleted orders within a finite time period.
3. The order hierarchical sorting and processing scheduling method considering machine maintenance according to claim 1, characterized in that, The maintenance time and node insertion sequence mentioned in step 5 are constructed in the following way: Initialize a sequence with the same length as the sub-order processing sequence, each position of the sequence corresponds to the process number of the sub-order processing sequence, and judge the value of the health value evaluation function of the machine when processing the process set; If the value is greater than the pre-maintenance threshold, the machine continues to process, and the corresponding position of the sequence is marked with label 1. When the value of the machine health value evaluation function is less than or equal to the pre-maintenance threshold after a certain process is completed, the machine status at the beginning of the process is set to the unselectable state. When the pre-maintenance is completed, the machine status is set to the selectable state, and the corresponding position of the sequence is marked with label 2. If a machine enters fault repair after completing a certain process, and the machine's pre-maintenance count reaches a threshold, the machine's state when the previous process was completed is set to an unselectable state. Then, label 3 is marked at the corresponding position in the sequence, thus obtaining the maintenance time and node insertion sequence.
4. The order hierarchical sorting and processing scheduling method considering machine maintenance according to claim 1, characterized in that, The preparation time and node insertion sequence mentioned in step 5 are constructed in the following way: Initialize a sequence with the same length as the sub-order processing sequence, each position of the sequence corresponds to the process number of the sub-order processing sequence, and judge the failure rate of important machine components; If the failure rate of a critical component of the machine is lower than the failure rate threshold, the machine continues processing, and the corresponding position in the sequence is marked with label 1a. This continues until a certain process is completed and the failure rate of the critical component of the machine is higher than or equal to the failure rate threshold. Then, the following judgment and operation are performed: when the machine's idle time is greater than or equal to the preparation time, label 1a is marked in the corresponding position in the sequence; when the machine's idle time is less than the preparation time, preparation time needs to be inserted before workpiece processing, and label 1b is marked in the corresponding position in the sequence, thus obtaining the preparation time and node insertion sequence.