An adaptive dynamic scheduling method for parts manufacturing

By optimizing the production scheduling of aerospace manufacturing parts through MILP modeling and tabu search, the algorithm deadlock problem was solved, enabling rapid generation and dynamic adjustment, thereby improving equipment utilization and parts delivery rate.

CN123070577APending Publication Date: 2026-11-10CHENGDU AIRCRAFT INDUSTRY GROUP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202611267692.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-20
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing production scheduling models for parts manufacturing in the aerospace industry are prone to algorithm lockout, resulting in processing delays and low equipment utilization, making it impossible to respond quickly to order changes and sudden equipment failures.

Method used

By combining MILP precise modeling and tabu search, production planning is optimized through adaptive tabu strategies. By defining decision variables and objective functions, production scheduling is optimized, conflicts are detected and resolved, and rapid generation and dynamic adjustment are achieved.

Benefits of technology

It improved equipment utilization by more than 10% and on-time parts delivery rate by more than 15%, ensuring the feasibility and flexibility of production plans and avoiding algorithm lock-in and resource waste.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123070577A_ABST
    Figure CN123070577A_ABST
Patent Text Reader

Abstract

This invention discloses an adaptive dynamic scheduling method for parts manufacturing, relating to the fields of production scheduling and aerospace manufacturing technology. The method includes problem modeling and MILP relaxation solution. First, a mixed integer programming model containing constraints such as parts demand, machine capacity, and production cost is constructed. By relaxing integer variables, it is transformed into a linear programming problem, which is solved quickly to obtain continuous solutions. Adaptive tabu search is used to avoid repeated searches while balancing global exploration and local optimization. Real-time conflict detection and repair are implemented, a multi-level conflict detection mechanism is designed, and a dynamic priority strategy is combined to automatically generate repair solutions, ensuring the feasibility of the scheduling plan.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of production scheduling and aerospace manufacturing technology, specifically to an adaptive dynamic scheduling method for parts manufacturing. Background Technology

[0002] Conventional aerospace manufacturing involves the production scheduling of a large number of parts. Before production scheduling, a pre-designed data model is used to conduct business planning for a large number of parts. The business planning process requires a settling period to pre-set the scheduling decisions. Even if a single heuristic algorithm is added to the data model, the data model will often experience algorithm lock-up in the local part where the single heuristic algorithm is added during the production scheduling process. This requires frequent testing and manual correction of the local part of the single heuristic algorithm, and at the same time, the output of the data model is no longer a complete production schedule.

[0003] Algorithm deadlocks during model execution are all triggered by factors relying on a single rule. However, the problem of algorithmic deadlocks essentially depends on the processing order of the algorithm itself. Therefore, the real cause of recurring deadlocks lies in the irreversibility of the processing order at the algorithm design level.

[0004] Current data models, in pursuit of transparency and directness between results and problems, and the lowest cost under the model, basically require pre-setting of production scheduling decisions. This leads to significant delays in handling changes in order demand and sudden equipment failures, and these two issues cannot be quantified during the design phase.

[0005] Chinese patent application CN202211679109.3, published on August 18, 2023, with publication number CN116611624A, outlines and refines existing processes and workflows. It employs advanced scheduling concepts and algorithms for scientific analysis, identifying bottlenecks on the production line and optimizing resource allocation through algorithmic optimization. This results in a rational arrangement of production plans, maximizing the capacity of key processes (equipment). The patent meets the demands of production lines with diverse product varieties, small batches, and frequent plan changes. By analyzing the matching between production plans and progress in real time, and automatically calculating and scheduling tasks in case of anomalies, it minimizes the impact of production changes on composite material production plans.

[0006] The integer programming model disclosed in this patent document creates a heuristic algorithm within the integer programming model to handle unexpected problems. This heuristic algorithm is used to solve the production scheduling problem, and greedy filling is used to optimize the efficiency of handling unexpected problems. Although this improves the effectiveness of integer programming modeling in handling unexpected problems, the algorithm is still prone to deadlock due to the irreversible processing order and the fixed local optimization caused by greedy filling, which can affect the progress of the production schedule. Summary of the Invention

[0007] To overcome the shortcomings of the prior art, this invention provides an adaptive dynamic scheduling method for parts manufacturing. By combining the accurate modeling capabilities of MILP with the efficient local optimization of tabu search, this invention enables the rapid generation and dynamic adjustment of production plans, ensuring the feasibility of the plan and improving equipment utilization and on-time delivery rate of parts.

[0008] To achieve the above-mentioned objectives, the technical solution of the present invention is as follows: An adaptive dynamic scheduling method for parts manufacturing includes the following steps: S1. Receive production data and construct a MILP model from the acquired part attributes, part manufacturing machinery capacity, and part manufacturing machinery usage status. S2. Based on the MILP model established in S1, calculate the defined decision variables for part scheduling. Use the defined decision variables to analyze and match the objective function for minimizing the total cost. At the same time, determine the constraints based on the part attributes, part manufacturing machinery capacity, and part manufacturing machinery usage status. Combine the defined decision variables, objective function, and constraints to calculate and obtain the range of values ​​for model relaxation and solution, which is used to complete the division of the optimal region. S3. Optimize the model relaxation and the specific numerical values ​​of the solution. Based on these values, divide the number of machine tools for part manufacturing and reorder the production sequence of parts. At the same time, optimize the production sequence and the division of machine tools for part manufacturing through an adaptive taboo strategy, and detect production conflicts. If a scheduling conflict is detected during the process of dividing the number of manufacturing machines for parts and reordering the production sequence of parts, resource conflict detection and time conflict detection are used to determine the conflict status. After the conflict status is confirmed, a conflict resolution strategy is used to redistribute the parts.

[0009] By adopting the above technical solutions, the production plan can be quickly generated and dynamically adjusted by combining the accurate modeling capabilities of MILP with the efficient local optimization of tabu search, ensuring the feasibility of the solution and improving equipment utilization and on-time delivery rate of parts.

[0010] The present invention is further configured such that: the part attributes include delivery time, priority, and material requirements, used to determine the requirements of each part; the part manufacturing machinery capacity includes the maximum capacity of the available machinery per unit time, used to calculate the maximum work done by the machinery during the delivery time; the part manufacturing machinery usage status includes the current usage status and usable performance of the machinery, used to obtain the specific status of the machinery during production scheduling.

[0011] By adopting the above technical solutions, the specific priority and delivery time of parts can be accurately obtained, and the time-limited validity information of parts can be confirmed. The maximum capacity of the parts manufacturing machinery per unit time and the usage status of the parts manufacturing machinery, including the current usage status and usable performance of the machinery, can be obtained to obtain the specific status of the machinery during production scheduling, and to confirm the effective dynamic time-limited information of the parts manufacturing machinery. In other words, the construction of the MILP model is based on the accurate acquisition of parts with production demand and the effective dynamic status of machinery, providing effective data for subsequent steps.

[0012] The present invention is further configured such that: in S2, decision variables are defined using data from the MILP model to determine whether the availability of machinery for scheduling parts within the delivery time is met, and these variables are obtained through Equations 1 and 2, respectively. Equation 1 is: Equation 1 (∈{0,1}) in, To indicate whether the machinery available within the delivery time can meet the demand for production parts, 0 means no and 1 means yes. Equation 2 is: ∈Z + Formula 2.

[0013] in, Z is the quantity of a unit part that can be produced on available machinery within the delivery timeframe. + It is a set of non-negative integers.

[0014] By adopting the above technical solution, it is possible to determine whether the parts to be scheduled meet the basic scheduling conditions, which is used to judge whether the dynamically acquired part data and manufacturing machinery data in the MILP model are compatible, and to make a preliminary screening of the completion rate of the parts to be scheduled.

[0015] The present invention is further configured such that: in S2, the objective function is determined using a pre-designed cost estimate, and whether the objective function meets the requirements is obtained through Equation 3. Equation 3 is: Formula 3 Where N is the projected cost. For production costs, To reduce replacement costs, For the delay time of the part, To delay punishment.

[0016] By adopting the above technical solution, Equation 3 is used to determine whether the production schedule parts meet the pre-designed cost estimate, in order to prevent losses from occurring directly during production scheduling. The pre-designed cost estimate is set based on the minimum requirement of the order price. Different orders have different calculation methods, and specific analysis and calculation are required based on the specific order situation. This is used to determine whether the production cost of the parts that need to be scheduled in the MILP model meets the basic production scheduling conditions and further confirms whether the basic production cost standard has been reached.

[0017] The present invention is further configured such that: the constraint conditions in S2 are used to determine whether the current part meets the production scheduling conditions through Equation 4. Equation 4 is: Formula 4 in, The unit time spent on the machine for a part. This represents the total available capacity of the machine. Let U be the total demand for the parts, and U be the maximum demand for the parts.

[0018] By adopting the above technical solution, it is determined whether the main problems encountered during the production of the scheduled parts are met, which is used to determine whether the basic conditions for the scheduled parts in the MILP model meet the constraints.

[0019] The present invention is further configured such that: in S2, model relaxation and solution relax integer variables into continuous variables, specifically 0 ≤ ≤ ,in, These are the numerical values ​​of each variable in continuous space. j It is a set of non-negative integers, derived from system order data.

[0020] By adopting the above technical solution, the optimal values ​​of each variable in continuous space are obtained, thereby realizing the division of areas where production scheduling of parts can be solved.

[0021] The present invention is further configured as follows: In S3, the optimal value of each variable in the continuous space is rounded down, and the specific calculation is obtained through Equation 5. Equation 5 is: Formula 5 in, For integer variables, These are the optimal values ​​of each variable in continuous space.

[0022] By adopting the above technical solution, the optimal values ​​of each variable in the continuous space are further optimized, the number of traversals is reduced, and the computational effect is improved.

[0023] The invention is further configured as follows: In S3, the adaptive tabu strategy divides the specific tabu search range by optimizing the model relaxation and solution values. During tabu search, a tabu range is provided to enable multiple visits to local optima. When the region is far from the optimal region, it is necessary to return to S2 to dynamically divide the region again and skip that region, thus limiting the potential optimal solution by tabu restrictions. When the region is close to the optimal value, the local fine search is strengthened and the structure of the current high-quality solution is protected by increasing the tabu range. The dynamic tabu length is calculated using Equation 6. Equation 6 is: Formula 6 Where L is the dynamic taboo length. The basic taboo length has a value range of [5, 15]. The objective value of the current solution. α is the target value of the historical best solution, and α is the adjustment coefficient.

[0024] By adopting the above technical solution, the optimal values ​​of each optimized variable in the continuous space can be traversed.

[0025] The present invention is further configured such that: when dynamically dividing regions, an iterative improvement strategy is used to judge the divided regions; Stop the search when the optimal value is found; if the improvement decreases, increase the taboo range; if the value is far from the optimal region, shorten the taboo range.

[0026] By adopting the above technical solution, the search range and space of the optimization area are automatically adjusted.

[0027] The present invention is further configured as follows: the iterative improvement strategy generates N candidate neighborhood solutions in each iteration, while setting a dynamic acceptance threshold and allowing a limited range of solution quality degradation. The range of the N candidate neighborhood solutions is [10, 30], while the range of the dynamic acceptance threshold is [1.03, 1.08], and the range of allowed solution quality degradation is [3%, 8%]. If no solution with a limited acceptable range of quality degradation is found, iteration stops after the maximum number of iterations is reached (150, 400). If the optimal solution is not improved after P consecutive iterations, iteration stops after P iterations (25, 50). If the computation time exceeds a preset threshold, iteration stops when the time exceeds the preset threshold (30, 300 seconds).

[0028] By adopting the above technical solution, the adjusted search partition region is traversed, and the traversal is constrained to prevent the traversal region from entering an infinite loop or from being unevenly distributed, which would cause the sequential traversal loop to be too long and result in a dead operation.

[0029] The present invention is further configured such that: if a production scheduling conflict is detected during the dynamic adjustment process, it is determined whether the manufacturing production scheduling conflict satisfies resource conflict detection or time conflict detection, or both resource conflict detection and time conflict detection are satisfied, and then resources are reallocated according to part priority using a conflict resolution strategy. Resource conflict detection is obtained through Equation 7, which is as follows: Formula 7 in, This refers to the resource requirements for the parts. For the total amount of resources, For machine i production The total amount of resource k occupied by part j, including tooling, fixtures and jigs, personnel qualifications, measuring tools and testing equipment, workstations and workbenches; Time conflict detection is obtained through Equation 8. Formula 8 is as follows: Formula 8 in, The start time of the part. For processing time, This refers to the deadline for parts delivery.

[0030] The above technical solution is used to determine the type of production scheduling conflict detected during the dynamic adjustment process.

[0031] The present invention is further configured such that: the conflict resolution strategy reallocates resources based on part priority, adjusts priorities, and obtains the result through dynamic priority calculation using Equation 9. Formula 9 is as follows: Formula 9 in, Priority The value of the component is w, and the weighting coefficient is w. Three components, of which The value range is [0.2, 0.5]. The value range is [0.3, 0.6]. The value range is [0.1, 0.3]. Assign importance to parts. Due to the urgency of time, The percentage of economic value is less than or equal to 0.3.

[0032] By adopting the above technical solution, the priority of parts can be readjusted.

[0033] In summary, the present invention has the following beneficial effects: 1. Compared with existing technologies, this invention can effectively combine the accurate modeling capabilities of MILP with the efficient local optimization of tabu search to achieve rapid generation and dynamic adjustment of production plans, ensuring the feasibility of the plan, improving equipment utilization and on-time delivery rate of parts, with equipment utilization increasing by more than 10% and on-time delivery rate of parts increasing by more than 15%.

[0034] 2. This invention constructs a MILP model by combining part attributes, manufacturing machinery capacity, and manufacturing machinery usage status. This model obtains the available status and maximum capacity of scheduled parts and available manufacturing machinery, enabling precise identification of the real-time status of parts and manufacturing machinery. By defining decision variables, objective functions, and constraints, the model determines the fit of imported parts and manufacturing machinery, thereby determining whether the scheduled order parts meet the technical production requirements. Simultaneously, all compatible scheduled order parts and manufacturing machinery are divided into multiple regions, achieving a preliminary division of all potentially compatible parts and manufacturing machinery within the factory area. This avoids omissions and significantly improves the screening basis for fit, preventing the need for a detailed search of all manufacturing machinery as soon as a scheduled order part appears. This reduces computation time and improves computational fit efficiency.

[0035] 3. This invention optimizes the partitioning space by rounding down incomplete production parts within the partitioned space, thereby achieving further optimization of the search. It utilizes tabu search to traverse each partitioned space once, and optimizes the traversal during the traversal process. The search stops when the optimal value is found, the improvement margin is reduced, the tabu range is increased, the area is moved away from the optimal region, and the tabu range is shortened, thereby further improving the search efficiency.

[0036] 4. This invention adjusts the search space by iteratively improving the strategy, limiting the search process to avoid dead loops. If the search space does not meet the adjustment requirements, it enters the next space. During the search process, scheduling conflicts need to be detected to avoid resource and time conflicts in the production order parts. If a conflict occurs, resources are reallocated using priority to avoid locking or skipping the production order parts.

[0037] 5. In this invention, the scheme is allocated according to the location of the manufacturing machinery with fixed design in the system-level neighborhood and the location of the parts in the production order. The parts are transferred using the machine-level neighborhood, resulting in waste of parts in the production scheduling and an increase in transportation time. Attached Figure Description

[0038] Figure 1This is a schematic diagram of the MILP relaxation solution process of the present invention; Figure 2 This is a schematic diagram of the tabu search optimization process of the present invention. Detailed Implementation

[0039] The specific implementation of the present invention will be described below with reference to the embodiments. Of course, the scope of protection of the present invention is not limited to the following embodiments.

[0040] Example 1 Refer to the instruction manual appendix Figure 1 and Figure 2 This embodiment discloses an adaptive dynamic scheduling method for parts manufacturing, including the following steps: S1, receiving production data and constructing a MILP model from the acquired parts attributes, parts manufacturing machine capacity, and parts manufacturing machine usage status; S2, calculating the defined decision variables for parts scheduling based on the MILP model established in S1, using the defined decision variables to analyze and match the objective function for minimizing total cost, and determining constraints based on the parts attributes, parts manufacturing machine capacity, and parts manufacturing machine usage status, and calculating the range of values ​​for model relaxation and solution based on the defined decision variables, objective function, and constraints, to complete the division of the optimal region; S3, optimizing the specific values ​​of model relaxation and solution, dividing the number of parts manufacturing machines based on these values, completing the reordering of parts production sequence, and simultaneously optimizing the production sequence and the division of the number of parts manufacturing machines through an adaptive tabu strategy, and detecting scheduling conflicts; if a scheduling conflict is detected during the reordering of the number of parts manufacturing machines and the parts production sequence, the conflict status is judged using resource conflict detection and time conflict detection, and after the conflict status is confirmed, the conflict is redistributed using a conflict resolution strategy.

[0041] In this embodiment, by effectively combining the precise modeling capabilities of MILP with the efficient local optimization of tabu search, the production plan can be generated and dynamically adjusted quickly, ensuring the feasibility of the solution and improving equipment utilization and on-time delivery rate of parts. Specifically, equipment utilization is increased by more than 10%, and on-time delivery rate of parts is increased by more than 15%.

[0042] Example 2 Compared with Example 1, the difference in this embodiment is that the part attributes include delivery time, priority, and material requirements, which are used to determine the requirements of each part; the part manufacturing machinery capacity includes the maximum capacity of available machinery per unit time, which is used to calculate the maximum work done by the machinery during the delivery time; the part manufacturing machinery usage status includes the current usage status and usable performance of the machinery, which is used to obtain the specific status of the machinery during production scheduling.

[0043] In S2, decision variables are defined using data from the MILP model to determine whether the availability of machinery for scheduling parts within the delivery time is met. These variables are obtained through Equations 1 and 2, respectively. Equation 1 is: Equation 1 ∈ {0,1} in, To indicate whether the machinery available within the delivery time can meet the demand for production parts, 0 means no and 1 means yes. Equation 2 is: ∈ Z + Formula 2.

[0044] in, Z is the quantity of a unit part that can be produced on available machinery within the delivery timeframe. + It is a set of non-negative integers.

[0045] In S2, the objective function is determined using a pre-designed cost estimate. The result is obtained through Equation 3, which is: Formula 3 Where N is the projected cost. For production costs, To reduce replacement costs, For the delay time of the part, To delay punishment.

[0046] In S2, the constraint condition is used to determine whether the current part meets the production scheduling conditions through Equation 4. Equation 3 is: Formula 4 in, The unit time spent on the machine for a part. This represents the total available capacity of the machine. Let U be the total demand for the parts, and U be the maximum demand for the parts.

[0047] In S2, model relaxation and solution involve relaxing integer variables into continuous variables, specifically 0 ≤ ≤ ,in, These are the numerical values ​​of each variable in continuous space. The set of non-negative integers is derived from the system's order data. By solving a relaxed linear programming problem, which includes decision variables, objective function (minimizing total cost), and constraints, the optimal values ​​of each variable in the continuous space are obtained.

[0048] Example 3 Compared with Example 2, the difference in this embodiment is that in this embodiment, the optimal value of each variable in the continuous space is rounded down in S3. The specific calculation is performed using Equation 5, which is: Formula 5 in, For integer variables, These are the optimal values ​​of each variable in continuous space.

[0049] In S3, the number of available part manufacturing machines is determined using machine-level neighborhood operations and operating system-level neighborhood operations to transfer parts from one machine to another and simultaneously adjust the allocation scheme of multiple parts. The reordering of part production sequence is accomplished using machine-level neighborhood operations. The production sequence of two parts on the same machine is swapped, and a single part is inserted into different positions on the same machine. The transfer and planning of parts and manufacturing machines are realized through system-level neighborhood and machine-level neighborhood, respectively. The system-level neighborhood is used to allocate adaptation schemes based on the distance between parts and manufacturing machines, and the transfer of parts is realized through machine-level neighborhood.

[0050] In S3, the adaptive tabu strategy optimizes the model relaxation and the specific numerical values ​​of the solution to define a specific tabu search range. During tabu search, a tabu range is provided to allow multiple visits to local optima. When the search moves far from the optimal region, it is necessary to return to S2 to dynamically redivide the region and skip it, thus limiting the potential optimal solution by preventing the tabu from being missed. Therefore, the tabu range needs to be shortened. When approaching the optimal value, to prevent oscillations near the optimal solution, strengthen the local fine search, and protect the structure of the current high-quality solution, the tabu range needs to be increased. The dynamic tabu length is calculated using Equation 6, which is: Formula 6 Where L is the dynamic taboo length. The basic taboo length has a value range of [5, 15]. The objective value of the current solution. α is the target value of the historical best solution, and α is the adjustment coefficient.

[0051] When dynamically dividing regions, an iterative improvement strategy is used to evaluate the divided regions. Stop the search when the optimal value is found; if the improvement decreases, increase the taboo range; if the value is far from the optimal region, shorten the taboo range.

[0052] The iterative improvement strategy generates N candidate neighborhood solutions in each iteration, while setting a dynamic acceptance threshold and allowing a limited range of solution quality degradation. The range of the N candidate neighborhood solutions is [10, 30], the dynamic acceptance threshold range is [1.03, 1.08], and the allowed range of solution quality degradation is [3%, 8%]. If no solution with a limited acceptable range of quality degradation is found, iteration stops after the maximum number of iterations is reached (150, 400). If the optimal solution is not improved after P consecutive iterations, iteration stops after P iterations (25, 50). If the computation time exceeds a preset threshold, iteration stops when the time exceeds the preset threshold (30, 300 seconds).

[0053] Example 4 Compared with Example 3, the difference in this embodiment is that, in this embodiment, if a production scheduling conflict is detected during the dynamic adjustment process, it is determined whether the manufacturing production scheduling conflict meets the requirements of resource conflict detection, time conflict detection, or both resource conflict detection and time conflict detection. Then, a conflict resolution strategy is used to reallocate resources according to the priority of the parts. Resource conflict detection is obtained through Equation 7, which is as follows: Formula 7 in, This refers to the resource requirements for the parts. For the total amount of resources, For machine i production The total amount of resource k occupied by part j, including tooling, fixtures, personnel qualifications, measuring tools and testing equipment, and workstations and workbenches.

[0054] Time conflict detection is obtained through Equation 8. Formula 8 is as follows: Formula 8 in, The start time of the part. For processing time, This refers to the deadline for parts delivery.

[0055] The conflict resolution strategy reallocates resources based on part priority, adjusts priorities, and uses dynamic priority calculation, as shown in Equation 9: Formula 9 in, Priority The value of the component is w, and the weighting coefficient is w. Three components, of which The value range is [0.2, 0.5]. The value range is [0.3, 0.6]. The value range is [0.1, 0.3]. Assign importance to parts. Due to the urgency of time, The proportion of economic value shall not exceed 0.3.

[0056] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. An adaptive dynamic scheduling method for parts manufacturing, characterized in that: Includes the following steps, S1. Receive production data and construct a MILP model from the acquired part attributes, part manufacturing machinery capacity, and part manufacturing machinery usage status. S2. Based on the MILP model established in S1, calculate the defined decision variables for part scheduling. Use the defined decision variables to analyze and match the objective function for minimizing the total cost. At the same time, determine the constraints based on the part attributes, part manufacturing machinery capacity, and part manufacturing machinery usage status. Combine the defined decision variables, objective function, and constraints to calculate and obtain the range of values ​​for model relaxation and solution, which is used to complete the division of the optimal region. S3. Optimize the model relaxation and the specific numerical values ​​of the solution. Based on these values, divide the number of machine tools for part manufacturing and reorder the production sequence of parts. At the same time, optimize the production sequence and the division of machine tools for part manufacturing through an adaptive taboo strategy, and detect production conflicts. If a scheduling conflict is detected during the process of dividing the number of manufacturing machines for parts and reordering the production sequence of parts, resource conflict detection and time conflict detection are used to determine the conflict status. After the conflict status is confirmed, a conflict resolution strategy is used to redistribute the parts.

2. The adaptive dynamic scheduling method for parts manufacturing according to claim 1, characterized in that: Part attributes include delivery time, priority, and material requirements, used to determine the demand for each part; part manufacturing machinery capacity includes the maximum capacity of available machinery per unit time, used to calculate the maximum work done by the machinery within the delivery time; part manufacturing machinery usage status includes the current usage status and available performance of the machinery, used to obtain the specific status of the machinery during production scheduling.

3. The adaptive dynamic scheduling method for parts manufacturing according to claim 2, characterized in that: In S2, decision variables are defined using data from the MILP model to determine whether the availability of machinery for scheduling parts within the delivery time is met. These variables are obtained through Equations 1 and 2, respectively. Equation 1 is: Equation 1 (∈{0,1}) in, To indicate whether the machinery available within the delivery time can meet the demand for production parts, 0 means no and 1 means yes. Equation 2 is: ∈Z + Formula 2. in, Z is the quantity of a unit part that can be produced on available machinery within the delivery timeframe. + It is a set of non-negative integers.

4. The adaptive dynamic scheduling method for parts manufacturing according to claim 2, characterized in that: In S2, the objective function is determined using a pre-designed cost estimate. The result is obtained via Equation 3. Equation 3 is: Formula 3 Where N is the projected cost. For production costs, To reduce replacement costs, For the delay time of the part, To delay punishment.

5. The adaptive dynamic scheduling method for parts manufacturing according to claim 2, characterized in that: In S2, the constraint conditions are used to determine whether the current part meets the production scheduling conditions through Equation 4. Equation 4 is: Formula 4 in, The unit time spent on the machine for a part. This represents the total available capacity of the machine. Let U be the total demand for the parts, and U be the maximum demand for the parts.

6. The adaptive dynamic scheduling method for parts manufacturing according to claim 5, characterized in that: In S2, model relaxation and solution involve relaxing integer variables into continuous variables, specifically 0 ≤ ≤ ,in, These are the numerical values ​​of each variable in continuous space. It is a set of non-negative integers, derived from system order data.

7. The adaptive dynamic scheduling method for parts manufacturing according to claim 2, characterized in that: In S3, the optimal values ​​of each variable in the continuous space are rounded down. The specific calculation is obtained through Equation 5. Equation 5 is: Formula 5 in, For integer variables, These are the optimal values ​​of each variable in continuous space.

8. The adaptive dynamic scheduling method for parts manufacturing according to claim 2, characterized in that: In S3, the adaptive tabu strategy optimizes the model relaxation and the specific numerical values ​​of the solution to define a specific tabu search range. During tabu search, a tabu range is provided to allow multiple visits to local optima. When the search moves far from the optimal region, it returns to S2 to dynamically redivide the region and skip it, thus limiting potential optimal solutions by preventing them from being missed. When approaching the optimal value, the search is strengthened locally to refine the search and protect the structure of current high-quality solutions, increasing the tabu range. The dynamic tabu length is calculated using Equation 6. Equation 6 is: Formula 6 Where L is the dynamic taboo length. The basic taboo length has a value range of [5, 15]. The objective value of the current solution. α is the target value of the historical best solution, and α is the adjustment coefficient.

9. The adaptive dynamic scheduling method for parts manufacturing according to claim 8, characterized in that: When dynamically dividing regions, an iterative improvement strategy is used to evaluate the divided regions. Stop the search when the optimal value is found; if the improvement decreases, increase the taboo range; if the value is far from the optimal region, shorten the taboo range.

10. The adaptive dynamic scheduling method for parts manufacturing according to claim 9, characterized in that: The iterative improvement strategy generates N candidate neighborhood solutions in each iteration, while setting a dynamic acceptance threshold and allowing a limited range of solution quality degradation. The range of the N candidate neighborhood solutions is [10, 30], the range of the dynamic acceptance threshold is [1.03, 1.08], and the range of allowed solution quality degradation is [3%, 8%]. If no solution with a limited quality degradation range is found, the iteration stops after the maximum number of iterations is reached, which is [150, 400]. If the optimal solution is not improved after P consecutive iterations, the iteration is stopped after P iterations. The range of P iterations is [25, 50]. If the calculation time exceeds a preset threshold, the iteration will stop when the time exceeds the preset threshold. The preset threshold range is [30, 300] seconds.

11. The adaptive dynamic scheduling method for parts manufacturing according to claim 10, characterized in that: If a production scheduling conflict is detected during the dynamic adjustment process, it is determined whether the manufacturing scheduling conflict meets the requirements of resource conflict detection, time conflict detection, or both. If so, a conflict resolution strategy is used to reallocate resources based on part priority. Resource conflict detection is obtained through Equation 7, which is as follows: Formula 7 in, This refers to the resource requirements for the parts. For the total amount of resources, For machine i production The total amount of resource k occupied by part j, including tooling, fixtures and jigs, personnel qualifications, measuring tools and testing equipment, workstations and workbenches; Time conflict detection is obtained through Equation 8. Formula 8 is as follows: Formula 8 in, The start time of the part. For processing time, This refers to the deadline for parts delivery.

12. The adaptive dynamic scheduling method for parts manufacturing according to claim 11, characterized in that: The conflict resolution strategy reallocates resources based on part priorities, adjusts priorities using dynamic priority calculation, and obtains the results via Equation 9. Formula 9 is as follows: Formula 9 in, Priority The value of the component is w, and the weighting coefficient is w. Three components, of which The value range is [0.2, 0.5]. The value range is [0.3, 0.6]. The value range is [0.1, 0.3]. Assign importance to parts. Due to the urgency of time, The percentage of economic value is less than or equal to 0.3.

Citation Information

Patent Citations

  • Intelligent production scheduling algorithm for composite parts

    CN116611624A