A Business Rule-Based Grouped Non-Equal Parallel Machine Scheduling Method
By using a business rule-based grouped non-equal parallel machine scheduling method, the scheduling scheme of the previous shift and machine qualification strategy are utilized to optimize the non-equal parallel machine scheduling in large-scale production. This solves the scheduling complexity and resource constraints in multi-variety, small-batch production, and improves production efficiency and continuity.
Patent Information
- Application Number
- CN202310701979.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-13
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2043-06-13
AI Technical Summary
Existing technologies struggle to effectively address the scheduling problem of non-parallel machines in large-scale production, especially in multi-variety, small-batch production. This presents challenges due to scheduling complexity, resource constraints, and multi-level limitations, making manual scheduling labor-intensive and difficult to meet production demands.
A business rule-based grouped non-equal parallel machine scheduling method is designed. By obtaining the scheduling scheme of the previous shift, the method generates a local scheduling scheme for the current shift using an inheritance strategy, and optimizes the scheduling process by combining machine qualifications and intensive production strategies.
It enables rapid optimization of scheduling within a reasonable timeframe, reduces the generation of illegal solutions, improves production continuity and intensive production efficiency, and reduces the workload of manual scheduling.
Smart Images

Figure CN116703098B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of parallel machine scheduling technology, and more specifically, to a grouped non-equal parallel machine scheduling method based on business rules. Background Technology
[0002] With economic globalization and the development of Industry 4.0, manufacturing has become the lifeline of national economic development in various countries. Most manufacturing industries are beginning to transform and upgrade towards intelligent manufacturing and smart factories. Due to the continuous emergence of personalized demands in the market environment, multi-variety, small-batch production has become an important production mode. As one of the most important decision-making processes in the manufacturing system, production scheduling is the key to the intelligent transformation and upgrading of enterprises. How to achieve the optimal trade-off between large-scale continuous production and the diversified output demand of fragmented orders is the key to the intelligent transformation and upgrading of enterprises. The non-equal parallel machine scheduling problem is a special kind of shop floor scheduling problem, which is widely used in manufacturing enterprises. Previous research on resource constraints in non-equal parallel machine scheduling is relatively limited. Most of the research focuses on resource-constrained project scheduling problems, and most of them remain at the theoretical level with little practical engineering application. In the limited parallel machine scheduling problems with mold constraints, they are mainly small and medium scale. At the same time, existing research has even less research on algorithms for grouping parallel machines by type. In addition, in the actual production process, there are not only various equipment constraints, such as sequentially dependent equipment adjustment, machine qualifications, and equipment quantity limits, but also processing constraints at other stages, as well as many practical requirements such as work calendars.
[0003] In actual production, due to the large number of production orders and numerous constraints, the workload of manual scheduling is heavy, making it difficult to enhance the value of schedulers. Currently, relying solely on planners' manual decisions can no longer meet the needs of factory production scheduling, and relying solely on optimization algorithms for decision-making is far from the expectations of actual schedulers, both of which have limitations.
[0004] Although scheduling such problems is quite difficult, planners have developed some core principles over the years. They improve upon these core principles to conduct targeted grouped parallel machine production scheduling, achieving large-scale continuous production while meeting the diverse order demands of multiple varieties.
[0005] Therefore, this invention addresses the challenge of summarizing and generalizing the scheduling experience of enterprise planners, which makes it difficult to translate their intended goals and objectives into corresponding indicators, and further into constraints (rules) with parameter settings. It designs a heuristic algorithm that integrates the business rules of planners, and utilizes the enterprise's production big data to quickly optimize and solve scheduling problems. Summary of the Invention
[0006] The purpose of this invention is to overcome the challenges posed by the explosive increase in problem-solving thresholds due to the continuous expansion of production scale and the accumulation of orders. This makes it difficult for precise algorithms to provide optimal solutions within a reasonable timeframe. Furthermore, multi-dimensional and multi-level constraints make it difficult to solve problems using simple scheduling rules. Resource limitations during scheduling lead intelligent optimization algorithms to waste significant search time and generate numerous invalid solutions through random generation. This invention draws inspiration from manual production scheduling, which proceeds shift by shift, generating shift scheduling plans according to the work calendar. First, planners configure business rule parameters based on the order and workshop layout. Then, they obtain the previous... The shift scheduling plan calls the inheritance strategy based on the previous shift's scheduling plan. Under the constraints and business rules, it inherits the scheduling plan of the machines in the previous shift as much as possible as possible as the current scheduling plan. At this time, it checks whether there are any unscheduled machines. If there are, a scheduling strategy that considers the machine's qualifications is adopted to arrange production for the unscheduled machines. If not, the scheduling of the current shift is completed. Furthermore, in order to meet the customer's demand for intensive production of products, an enhancement strategy based on intensive production is designed to improve the intensive production of products. As the work calendar shifts progress, the construction of the refrigerator foaming workshop scheduling plan is completed.
[0007] To achieve the above objectives, the present invention provides the following technical solution:
[0008] A business rule-based grouped non-equal parallel machine scheduling method, characterized by the following steps:
[0009] S1: Obtain the scheduling plan for the previous shift;
[0010] S2: Call the scheduling scheme of the previous shift, and use the inheritance strategy based on the scheduling scheme of the previous shift to obtain the local scheduling scheme of the current shift.
[0011] S3: Determine whether there are any unscheduled machines based on the local scheduling plan for the current shift. If there are unscheduled machines, adopt a scheduling strategy that considers machine qualifications, arrange production for the unscheduled machines, and output the current shift plan. If there are no unscheduled machines, the final scheduling for the current shift is completed, and the current shift plan is output.
[0012] Furthermore, in step S2, the invocation of the previous shift's scheduling scheme, using an inheritance strategy based on the previous shift's scheduling scheme, includes the following steps:
[0013] S2.1: Invoke the scheduling scheme of the previous shift;
[0014] S2.2: Verify the continuous production days constraint, door supply capacity, and last order verification of the previous shift's scheduling plan;
[0015] S2.3: Obtain the local scheduling plan for the current shift.
[0016] Furthermore, in step S2.2, the verification of the continuous production days constraint, the inspection of door supply capacity, and the verification of the last order for the previous shift's scheduling plan specifically includes:
[0017] S2.21: Verify the continuous production days constraint for products with delayed production schedules. Products that have not reached the minimum continuous production days limit need to continue production.
[0018] S2.22: For products whose production schedule has exceeded the maximum continuous production days limit, determine whether the remaining orders for the product in the current month can be completed within the minimum continuous production days. If so, continue production; otherwise, process the products that have exceeded the maximum continuous production days limit by removing the mold.
[0019] S2.23: In conjunction with the work calendar, check the door supply capacity of the products, and perform mold processing for products with limited door supply capacity;
[0020] S2.24: In the event of leftover orders, the production quantity of the scheduled products will be reduced to obtain a partial scheduling plan for the current shift.
[0021] Furthermore, in step S3, the scheduling strategy that considers machine qualifications and arranges for unscheduled machines to produce specifically includes:
[0022] Step S3.01: Construct the set N of products that can be scheduled for production;
[0023] Step S3.02: Sort the set N of products that can be scheduled for production;
[0024] Step S3.03: Obtain the set of regions with qualified machines for products that can be scheduled for production;
[0025] Step S3.04: Determine if there are any idle machines in each region of the region set; if there are idle machines in each region of the region set, proceed to step S3.06; if there are no idle machines in each region of the region set, execute the product coordination rules.
[0026] Step S3.05: Update the scheduling information for each sub-region;
[0027] Step S3.06: Sort each region according to the number of idle machines, and select region P with the most idle machines for trial sorting;
[0028] Step S3.07: Determine whether the product meets the machine qualification requirements when produced in area P. Specifically: if the product meets the machine qualification requirements when produced in area P, calculate the mold type information in area P; if the product does not meet the machine qualification requirements when produced in area P, proceed to step S3.06 and try to arrange the product in the next area.
[0029] Step S3.08: Determine whether the number of mold types in region P has reached the maximum number of mold types and whether the molds required for the product are not scheduled for production in region P. Specifically: if the number of mold types in region P has reached the maximum number of mold types and the molds required for the product are not scheduled for production in region P, proceed to step S3.09; if the number of mold types in region P has not reached the maximum number of mold types and the molds required for the product are not scheduled for production in region P, proceed to step S3.06 and try scheduling in the next region.
[0030] Step S3.09: Collect information on idle devices within region P;
[0031] Step S3.10: Determine whether the number of idle machines in area P is greater than the number of products n to be tested. If yes, arrange all products n into area P. If not, arrange products n with the same number of machines into area P and prioritize processing them on machines that do not require mold replacement. Then proceed to step S3.13.
[0032] Step S3.11: Update the product layout quantity;
[0033] Step S3.12: Update machine information and mold type information for each area;
[0034] Step S3.13: Determine if there are still areas within the line that have not participated in scheduling. If yes, jump to step S3.09; otherwise, output the product scheduling result.
[0035] Furthermore, in step S3.01, the method for constructing the set N of products that can be scheduled for production is specifically as follows:
[0036] S3.011: Take the set of products with current order demand, and remove products in set N that have production mutual exclusion constraints with the current shift scheduling plan;
[0037] S3.012: Query the product production information table. Products that have just been removed from the mold will not participate in the upper mold. Based on this, further calculate the maximum number of products that can be arranged to obtain the set of products that can be arranged, N.
[0038] Furthermore, in step S3.04, the construction of the product coordination rules specifically involves:
[0039] S3.041: Sort products within region P according to their already scheduled production quantities;
[0040] S3.042: Prioritize adjusting the production scheduling area starting with the product n1, which has the lowest production quantity;
[0041] S3.043: Find the set of regions where the number of idle machines must be greater than or equal to the production quantity of product n1 and denote it as P, and sort them in order of priority according to the region with the most idle machines;
[0042] S3.044: Determine whether the product n1 exceeds the mold type constraint in area P1 after it is placed into partition P1. If it does, proceed to step S3.045; if it does not, proceed to step S3.046.
[0043] S3.045: Determine if there is an idle machine in area P1. If not, proceed to step S3.042 and try to arrange the next product. If there is, proceed to step S3.043 and try to arrange the next area.
[0044] S3.046: Move product n1 into area P1;
[0045] S3.047: Output area adjustment results;
[0046] Furthermore, in step S3.012, the maximum number of products that can be arranged is calculated as follows: maximum number of products that can be arranged = min(mold capacity, door mold capacity, door line fixture capacity) / daily production capacity of a single mold.
[0047] Furthermore, step S4 is included after step S3, where step S4 uses the enhancement strategy based on intensive production to optimize intensive production.
[0048] Furthermore, in step S4, the use of an enhancement strategy based on intensive production to optimize intensive production specifically includes:
[0049] Step S4.01: Calculate the set M of products M produced across regions in the current shift scheduling plan;
[0050] Step S4.02: Determine whether the product set M is an empty set. If it is an empty set, it means that there are no products produced across edges. Jump to step S4.17. If it is not an empty set, jump to step S4.03.
[0051] Step S4.03: Calculate the set P of all sorted regions for product n1;
[0052] Step S4.04: Based on the machine qualification matrix, obtain the set of regions l1 that can produce product n1;
[0053] Step S4.05: Sort each area according to the principle of prioritizing the largest number of mold types that can be handled and the smallest number of idle machines, and select the highest priority area P1 for trial adjustment;
[0054] Step S4.06: Obtain the set N of products that have been sorted in region P1, and take product n2 from set N;
[0055] Step S4.07: Obtain the set Q of all sorted regions for product n2;
[0056] Step S4.08: Sort the set in descending order according to the number of products n2 that have been sorted, and select the region P2 with the highest priority;
[0057] Step S4.09: Determine whether the production quantity of product n2 in P2 is greater than or equal to the production quantity of product n1 in P1. If yes, proceed to the next step; otherwise, jump to step S4.08 and try to arrange the next edge.
[0058] Step S4.10: Based on the machine qualification matrix, obtain the set of production areas l2 for product n2;
[0059] Step S4.11: Determine whether the product n2 is placed in region P1 and meets the machine qualification constraints. If region P1 belongs to the set of production regions l1, jump to step S4.12. If region P1 does not belong to the set of production regions l1, jump to step S4.08 and try to place the next edge.
[0060] Step S4.12: Determine whether all regions within Q have been tried and arranged. If yes, proceed to the next step. If no, jump to step S4.08 and try and arrange the next region.
[0061] Step S4.13: Determine whether all products within N have been tested. If yes, proceed to the next step; otherwise, jump to step S4.06 and test the next product.
[0062] Step S4.14: Determine whether all regions within P have been tried and arranged. If yes, proceed to the next step. If no, jump to step S4.05 and try and arrange the next region.
[0063] Step S4.15: Determine whether all products within M have been tested. If yes, output the adjustment result. If no, jump to step S4.03 and test the next product.
[0064] Step S4.16: Swap equal amounts of product n1 and product n2 regions;
[0065] Step S4.17: Determine whether the current plan is optimized compared to the original manually compiled plan. If the percentage of optimization is less than zero, proceed to step S4.08 and try the next area. If the percentage of optimization is greater than zero, output the adjustment result.
[0066] Furthermore, in step S4.17, determining whether the current solution is optimized compared to the original manually compiled solution specifically involves:
[0067]
[0068] f man For manually compiled scheme indicators, f rulef is a scheme metric for a business rule-based grouped non-equal parallel machine scheduling method. per The percentage of metrics optimized.
[0069] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:
[0070] Compared with existing scheduling methods, this invention addresses the challenge of summarizing and generalizing the scheduling experience of enterprise planners, identifying their intentions and objectives that are difficult to translate into corresponding indicators, and then transforming them into constraints with parameter settings. It designs a heuristic algorithm that integrates the business rules of planners, and utilizes the enterprise's production big data to quickly optimize and solve scheduling problems. Attached Figure Description
[0071] Figure 1 A flowchart of a grouped non-equal parallel machine scheduling method based on business rules provided by the present invention;
[0072] Figure 2 A flowchart of a grouped non-equal parallel machine scheduling method based on business rules is provided in an embodiment of the present invention;
[0073] Figure 3 A flowchart illustrating the method for the inheritance strategy based on the previous shift scheduling scheme provided in an embodiment of the present invention;
[0074] Figure 4 A flowchart of a scheduling strategy that considers machine qualifications, provided for an embodiment of the present invention;
[0075] Figure 5 A flowchart of product coordination rules provided for embodiments of the present invention;
[0076] Figure 6 A flowchart illustrating the enhancement strategy based on intensive production provided in this embodiment of the invention. Detailed Implementation
[0077] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.
[0078] To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions;
[0079] It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.
[0080] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.
[0081] Example 1
[0082] This embodiment provides a business rule-based grouped non-equal parallel machine scheduling method, such as... Figure 1 As shown, it includes the following steps:
[0083] S1: Obtain the scheduling plan for the previous shift;
[0084] S2: Call the scheduling scheme of the previous shift, and use the inheritance strategy based on the scheduling scheme of the previous shift to obtain the local scheduling scheme of the current shift.
[0085] S3: Based on the current shift's partial scheduling plan, determine if there are any unscheduled machines. If so, adopt a scheduling strategy that considers machine qualifications, arrange production for the unscheduled machines, and output the current shift's plan; if there are no unscheduled machines, the final scheduling for the current shift is complete, and the current shift's plan is output. Figure 2 As shown, in this embodiment, the business rule-based grouped non-equal parallel machine scheduling method draws on the shift-by-shift approach used in manual production scheduling. It generates shift scheduling plans according to the work calendar, shift by shift. First, the planner configures business rule parameters based on the order workshop layout. Then, it obtains the scheduling plan for the previous shift and calls the inheritance strategy based on the previous shift's scheduling plan. Under the constraints and business rule limitations, it inherits the scheduling plan of the machines from the previous shift as much as possible as possible as the current scheduling plan. At this point, it checks whether there are any unscheduled machines. If so, it adopts a scheduling strategy that considers machine qualifications and arranges production for the unscheduled machines. If not, the scheduling for the current shift is completed. Furthermore, to meet the customer's demand for intensive product production, an enhancement strategy based on intensive production is designed to improve the intensive production of products. As the work calendar shifts progress, the construction of the refrigerator foaming workshop scheduling plan is completed.
[0086] Example 2
[0087] This embodiment, based on Embodiment 1, continues to disclose the following content:
[0088] like Figure 3 As shown, the invocation of the previous shift's scheduling scheme, using an inheritance strategy based on the previous shift's scheduling scheme, includes the following steps:
[0089] S2.1: Invoke the scheduling scheme of the previous shift;
[0090] S2.2: Verify the continuous production days constraint, door supply capacity, and last order verification of the previous shift's scheduling plan;
[0091] S2.3: Obtain the local scheduling plan for the current shift.
[0092] The verification of the continuous production days constraint, door supply capacity, and last-minute order verification of the previous shift's scheduling plan specifically includes:
[0093] S2.21: Verify the continuous production days constraint for products with delayed production schedules. Products that have not reached the minimum continuous production days limit need to continue production.
[0094] S2.22: For products whose production schedule has exceeded the maximum continuous production days limit, determine whether the remaining orders for the product in the current month can be completed within the minimum continuous production days. If so, continue production; otherwise, process the products that have exceeded the maximum continuous production days limit by removing the mold.
[0095] S2.23: In conjunction with the work calendar, check the door supply capacity of the products, and perform mold processing for products with limited door supply capacity;
[0096] S2.24: In the event of leftover orders, the production quantity of the scheduled products will be reduced to obtain the local scheduling plan for the current shift.
[0097] Because the loading, unloading, and preheating of large molds require a long time, product production suffers from the problem of long mold switching times. However, large-scale mold changes during product production are unacceptable to planners in actual production. At the same time, due to the business rule of minimum continuous production days, product production needs a certain degree of continuity, meaning that scheduling the production of a certain product requires continuous production for a period of time. Since the previous shift's scheduling plan contains a large number of good segments, this strategy uses the previous shift's scheduling plan as a pre-scheduling plan, constrains and checks its resources, and uses the retained scheduling plan as the local scheduling plan for the current shift. This can not only effectively reduce mold switching and improve production continuity, but also reduce algorithm search time and improve algorithm search efficiency. In this embodiment, based on the fact that 15 types of machines were released in three areas A, B, and C in the previous shift, firstly, after constraint checks, unqualified machines 4, 5, and 6 in area B are removed. Then, after resource checks, three machines in area C are removed. Finally, after last-order checks, one unqualified machine is removed. The result is used as the local scheduling plan for the current shift.
[0098] The minimum continuous production days for a product refers to the period of continuous production required before the product can be removed from the mold. During scheduling, the minimum continuous production days set manually by the planner must be met. By limiting the minimum continuous production days, continuous production for a period of time is guaranteed, preventing the product from being removed from the mold due to conflicts with other soft constraints, thus improving the continuity of product production. The maximum continuous production days for a product refers to the time from production to removal from the mold, which cannot exceed a certain period of time. During scheduling, the maximum continuous production days set manually by the planner must be met. By limiting the maximum continuous production days, continuous production for a period of time is guaranteed.
[0099] Example 3
[0100] This embodiment, based on Embodiment 1, continues to disclose the following content:
[0101] like Figure 4 As shown, in step S3, the scheduling strategy that considers machine qualifications and arranges for unscheduled machines to produce specifically includes:
[0102] Step S3.01: Construct the set N of products that can be scheduled for production;
[0103] Step S3.02: Sort the set N of products that can be scheduled for production;
[0104] Step S3.03: Obtain the set of regions with qualified machines for products that can be scheduled for production;
[0105] Step S3.04: Determine if there are any idle machines in each region of the region set; if there are idle machines in each region of the region set, proceed to step S3.06; if there are no idle machines in each region of the region set, execute the product coordination rules.
[0106] Step S3.05: Update the scheduling information for each sub-region;
[0107] Step S3.06: Sort each region according to the number of idle machines, and select region P with the most idle machines for trial sorting;
[0108] Step S3.07: Determine whether the product meets the machine qualification requirements when produced in area P. Specifically: if the product meets the machine qualification requirements when produced in area P, calculate the mold type information in area P; if the product does not meet the machine qualification requirements when produced in area P, proceed to step S3.06 and try to arrange the product in the next area.
[0109] Step S3.08: Determine whether the number of mold types in region P has reached the maximum number of mold types and whether the molds required for the product are not scheduled for production in region P. Specifically: if the number of mold types in region P has reached the maximum number of mold types and the molds required for the product are not scheduled for production in region P, proceed to step S3.09; if the number of mold types in region P has not reached the maximum number of mold types and the molds required for the product are not scheduled for production in region P, proceed to step S3.06 and try scheduling in the next region.
[0110] Step S3.09: Collect information on idle devices within region P;
[0111] Step S3.10: Determine whether the number of idle machines in area P is greater than the number of products n to be tested. If yes, arrange all products n into area P. If not, arrange products n with the same number of machines into area P and prioritize processing them on machines that do not require mold replacement. Then proceed to step S3.13.
[0112] Step S3.11: Update the product layout quantity;
[0113] Step S3.12: Update machine information and mold type information for each area;
[0114] Step S3.13: Determine if there are still areas within the line that have not participated in scheduling. If yes, jump to step S3.09; otherwise, output the product scheduling result.
[0115] In step S3.01, the method for constructing the set N of products that can be scheduled for production is specifically as follows:
[0116] S3.011: Take the set of products with current order demand, and remove products in set N that have production mutual exclusion constraints with the current shift scheduling plan;
[0117] S3.012: Query the product production information table. Products that have just been removed from the mold will not participate in the upper mold. Based on this, further calculate the maximum number of products that can be arranged to obtain the set of products that can be arranged, N.
[0118] In step S3.01, the maximum number of products that can be arranged is calculated as follows: maximum number of products that can be arranged = min(mold capacity, door mold capacity, door line fixture capacity) / daily production capacity of a single mold.
[0119] Example 4
[0120] Based on Example 3, this embodiment continues to disclose the following content:
[0121] like Figure 5 As shown, in step S3.04, the construction of the product coordination rules specifically involves:
[0122] S3.041: Sort products within region P according to their already scheduled production quantities;
[0123] S3.042: Prioritize adjusting the production scheduling area starting with the product n1, which has the lowest production quantity;
[0124] S3.043: Find the set of regions where the number of idle machines must be greater than or equal to the production quantity of product n1 and denote it as P, and sort them in order of priority according to the region with the most idle machines;
[0125] S3.044: Determine whether the product n1 exceeds the mold type constraint in area P1 after it is placed into partition P1. If it does, proceed to step S3.045; if it does not, proceed to step S3.046.
[0126] S3.045: Determine if there is an idle machine in area P1. If not, proceed to step S3.042 and try to arrange the next product. If there is, proceed to step S3.043 and try to arrange the next area.
[0127] S3.046: Move product n1 into area P1;
[0128] S3.047: Output area adjustment results;
[0129] Example 5
[0130] Based on Examples 1 to 4, this embodiment continues to disclose the following content:
[0131] like Figure 6 As shown, step S3 is followed by step S4, which uses the enhancement strategy based on intensive production to optimize intensive production.
[0132] The aforementioned enhancement strategy based on intensive production to optimize intensive production specifically includes:
[0133] Step S4.01: Calculate the set M of products M produced across regions in the current shift scheduling plan;
[0134] Step S4.02: Determine whether the product set M is an empty set. If it is an empty set, it means that there are no products produced across edges. Jump to step S4.17. If it is not an empty set, jump to step S4.03.
[0135] Step S4.03: Calculate the set P of all sorted regions for product n1;
[0136] Step S4.04: Based on the machine qualification matrix, obtain the set of regions l1 that can produce product n1;
[0137] Step S4.05: Sort each area according to the principle of prioritizing the largest number of mold types that can be handled and the smallest number of idle machines, and select the highest priority area P1 for trial adjustment;
[0138] Step S4.06: Obtain the set N of products that have been sorted in region P1, and take product n2 from set N;
[0139] Step S4.07: Obtain the set Q of all sorted regions for product n2;
[0140] Step S4.08: Sort the set in descending order according to the number of products n2 that have been sorted, and select the region P2 with the highest priority;
[0141] Step S4.09: Determine whether the production quantity of product n2 in P2 is greater than or equal to the production quantity of product n1 in P1. If yes, proceed to the next step; otherwise, jump to step S4.08 and try to arrange the next edge.
[0142] Step S4.10: Based on the machine qualification matrix, obtain the set of production areas l2 for product n2;
[0143] Step S4.11: Determine whether the product n2 is placed in region P1 and meets the machine qualification constraints. If region P1 belongs to the set of production regions l1, jump to step S4.12. If region P1 does not belong to the set of production regions l1, jump to step S4.08 and try to place the next edge.
[0144] Step S4.12: Determine whether all regions within Q have been tried and arranged. If yes, proceed to the next step. If no, jump to step S4.08 and try and arrange the next region.
[0145] Step S4.13: Determine whether all products within N have been tested. If yes, proceed to the next step; otherwise, jump to step S4.06 and test the next product.
[0146] Step S4.14: Determine whether all regions within P have been tried and arranged. If yes, proceed to the next step. If no, jump to step S4.05 and try and arrange the next region.
[0147] Step S4.15: Determine whether all products within M have been tested. If yes, output the adjustment result. If no, jump to step S4.03 and test the next product.
[0148] Step S4.16: Swap equal amounts of product n1 and product n2 regions;
[0149] Step S4.17: Determine whether the current plan is optimized compared to the original manually compiled plan. If the percentage of optimization is less than zero, proceed to step S4.08 and try the next area. If the percentage of optimization is greater than zero, output the adjustment result.
[0150] In step S4.17, determining whether the current solution is optimized compared to the original manually compiled solution specifically involves:
[0151]
[0152] f man For manually compiled scheme indicators, f rule f is a scheme metric for a business rule-based grouped non-equal parallel machine scheduling method. per The percentage of metrics optimized.
[0153] In this embodiment, a survey was conducted on the actual production situation of the background enterprise's production plant. The order information of the enterprise's production line X49100 manually scheduled on August 10, 2022 was compiled. The production line of this factory required 14 types of products in August, and needed to use 6 types of molds with a total demand of 121,020 units. In order to ensure the security and privacy of the enterprise's data, and at the same time to display the data characteristics, some data was anonymized. The order information and the equipment information required for processing are shown in Table 1 below.
[0154] Table 1 Order Summary Information
[0155]
[0156]
[0157] The machine grouping information and machine qualification matrix are shown in Table 2 below.
[0158] Table 2 Line Information
[0159]
[0160]
[0161]
[0162] The evaluation of the scheme is carried out using indicators from the mathematical model, where the manually compiled scheme indicator is taken as f. man The scheme index f of the business rule-based grouped non-equal parallel machine scheduling method rule Calculate the percentage of index optimization f per The calculation formula is:
[0163]
[0164] The statistical indicators for the two schemes are shown in Table 3:
[0165] Table 3 Performance Index Information Table
[0166]
[0167] Table 4 shows the number of product types produced in the first eight cycles:
[0168] Table 4. Number of Product Types Produced in a Cycle
[0169]
[0170] Based on the data analysis above, comparing the manual scheduling scheme with the simulation scheduling results shows that the scheduling scheme based on heuristic algorithms is similar to the manually compiled scheme in terms of product diversity. However, the heuristic algorithm significantly improves the reduction of mold changes and intensive product production compared to the manually compiled scheme, proving the effectiveness of the heuristic algorithm based on business rules.
[0171] The same or similar labels correspond to the same or similar parts;
[0172] The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.
[0173] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.
Claims
1. A grouped non-equal parallel machine scheduling method based on business rules, characterized in that, Includes the following steps: S1: Obtain the scheduling plan for the previous shift; S2: Call the scheduling scheme of the previous shift, and use the inheritance strategy based on the scheduling scheme of the previous shift to obtain the local scheduling scheme of the current shift. S3: Determine whether there are any unscheduled machines based on the local scheduling plan for the current shift. If there are unscheduled machines, adopt a scheduling strategy that takes into account machine qualifications, arrange production for the unscheduled machines, and output the current shift plan. If there are no unscheduled machines, the final scheduling for the current shift is completed, and the current shift plan is output. In step S3, the scheduling strategy that considers machine qualifications and arranges for unscheduled machines to produce specifically includes: Step S3.01: Construct the set N of products that can be scheduled for production; Step S3.02: Sort the set N of products that can be scheduled for production; Step S3.03: Obtain the set of regions with qualified machines for products that can be scheduled for production; Step S3.04: Determine if there are any idle machines in each region of the region set; If there are idle machines in each region of the region set, proceed to step S3.06; if there are no idle machines in each region of the region set, execute the product coordination rules. Step S3.05: Update the scheduling information for each sub-region; Step S3.06: Sort each region according to the number of idle machines, and select region P with the most idle machines for trial sorting; Step S3.07: Determine whether the product meets the machine qualification requirements when produced in area P. Specifically: if the product meets the machine qualification requirements when produced in area P, calculate the mold type information in area P; if the product does not meet the machine qualification requirements when produced in area P, proceed to step S3.06 and try to arrange the product in the next area. Step S3.08: Determine whether the number of mold types in region P has reached the maximum number of mold types and whether the molds required for the product are not scheduled for production in region P. Specifically: if the number of mold types in region P has reached the maximum number of mold types and the molds required for the product are not scheduled for production in region P, proceed to step S3.09; if the number of mold types in region P has not reached the maximum number of mold types and the molds required for the product are not scheduled for production in region P, proceed to step S3.06 and try scheduling in the next region. Step S3.09: Collect information on idle devices within region P; Step S3.10: Determine whether the number of idle machines in area P is greater than the number of products n to be tested. If yes, arrange all products n into area P. If not, arrange products n with the same number of machines into area P and prioritize processing them on machines that do not require mold replacement. Then proceed to step S3.
13. Step S3.11: Update the product layout quantity; Step S3.12: Update machine information and mold type information for each area; Step S3.13: Determine if there are still areas within the line that have not participated in scheduling. If yes, jump to step S3.09; otherwise, output the product scheduling result.
2. The grouped non-equal parallel machine scheduling method based on business rules according to claim 1, characterized in that, In step S2, the invocation of the previous shift's scheduling scheme, using an inheritance strategy based on the previous shift's scheduling scheme, includes the following steps: S2.1: Invoke the scheduling scheme of the previous shift; S2.2: Verify the continuous production days constraint, door supply capacity, and last order verification of the previous shift's scheduling plan; S2.3: Obtain the local scheduling plan for the current shift.
3. The grouped non-equal parallel machine scheduling method based on business rules according to claim 2, characterized in that, In step S2.2, the verification of the continuous production days constraint, the inspection of door supply capacity, and the verification of the last order for the previous shift's scheduling plan specifically includes: S2.21: Verify the continuous production days constraint for products with delayed production schedules. Products that have not reached the minimum continuous production days limit need to continue production. S2.22: For products whose production schedule has exceeded the maximum continuous production days limit, determine whether the remaining orders for the product in the current month can be completed within the minimum continuous production days. If so, continue production; otherwise, process the products that have exceeded the maximum continuous production days limit by removing the mold. S2.23: In conjunction with the work calendar, check the door supply capacity of the products, and perform mold processing for products with limited door supply capacity; S2.24: In the event of leftover orders, the production quantity of the scheduled products will be reduced to obtain a partial scheduling plan for the current shift.
4. The grouped non-equal parallel machine scheduling method based on business rules according to claim 1, characterized in that, In step S3.01, the method for constructing the set N of products that can be scheduled for production is as follows: S3.011: Take the set of products with current order demand, and remove products in set N that have production mutual exclusion constraints with the current shift scheduling plan; S3.012: Query the product production information table. Products that have just been removed from the mold will not participate in the upper mold. Based on this, further calculate the maximum number of products that can be arranged to obtain the set of products that can be arranged, N.
5. The grouped non-equal parallel machine scheduling method based on business rules according to claim 1, characterized in that, In step S3.04, the construction of the product coordination rules specifically involves: S3.041: Sort products within region P according to their already scheduled production quantities; S3.042: Prioritize products with the lowest production quantities. Production scheduling area adjustments have begun; S3.043: The number of available machines must be greater than or equal to the number of products. The set of regions with the most production capacity is denoted as P, and the regions with the most idle machines are sorted first. S3.044: Determine when the product... Displace into partition Does it exceed the area? The internal mold type is constrained. If the number of types exceeds the limit, proceed to step S3.045; otherwise, proceed to step S3.
046. S3.045: Determine the region If there are any idle machines, proceed to step S3.042 and try scheduling the next product; if there are, proceed to step S3.043 and try scheduling the next area. S3.046: To make the product Move into area ; S3.047: Output area adjustment results.
6. The grouped non-equal parallel machine scheduling method based on business rules according to claim 4, characterized in that, In step S3.012, the maximum number of products that can be arranged is calculated as follows: maximum number of products that can be arranged = min(mold capacity, door mold capacity, door line fixture capacity) / daily production capacity of a single mold.
7. A business rule-based grouped non-equal parallel machine scheduling method according to any one of claims 1 to 6, characterized in that, Step S3 is followed by step S4, which uses an enhancement strategy based on intensive production to optimize intensive production.
8. The grouped non-equal parallel machine scheduling method based on business rules according to claim 7, characterized in that, In step S4, the use of an enhancement strategy based on intensive production to optimize intensive production specifically includes: Step S4.01: Calculate the set M of products M produced across regions according to the current shift scheduling plan; Step S4.02: Determine whether the product set M is an empty set. If it is an empty set, it means that there are no products produced across edges, and proceed to step S4.
17. If it is not an empty set, proceed to step S4.
03. Step S4.03: Statistical analysis of products The set P of all sorted regions; Step S4.04: Obtain product information based on the machine qualification matrix. Set of Producible Areas ; Step S4.05: Sort each region according to the principle of prioritizing the region with the largest number of mold types it can handle and the region with the smallest number of idle machines, and select the region with the highest priority. Perform trial adjustments; Step S4.06: Obtain the region Given a set N of products, retrieve the products within set N. ; Step S4.07: Obtain the product Q is the set of all sorted regions; Step S4.08: Assemble according to product Sort the already sorted quantities in descending order and select the region with the highest priority. ; Step S4.09: Determine the product exist Is the production quantity greater than or equal to the product quantity? exist If the production quantity is specified, proceed to the next step; otherwise, jump to step S4.08 and try scheduling on the next edge. Step S4.10: Obtain products based on the machine qualification matrix. Set of Producible Areas ; Step S4.11: Determine the product Discharge area Does it meet the machine qualification requirements? If the region Belongs to the set of productive regions Proceed to step S4.12, if the region Proceed to step S4.08 and try arranging the next edge; Step S4.12: Determine whether all regions within Q have been tried and sorted. If yes, proceed to the next step; otherwise, jump to step S4.08 and try sorting the next region. Step S4.13: Determine whether all products within combination N have been tested. If yes, proceed to the next step; otherwise, jump to step S4.06 and test the next product. Step S4.14: Determine whether all regions within P have been tried and sorted. If yes, proceed to the next step; otherwise, jump to step S4.05 and try and sort the next region. Step S4.15: Determine whether all products within M have been trial-arranged. If yes, output the adjustment result; otherwise, jump to step S4.03 and try the next product. Step S4.16: Take equal amounts of products With products Area swap; Step S4.17: Determine whether the current plan is optimized compared to the original manually compiled plan. If the percentage of optimization is less than zero, proceed to step S4.08 and try the next area. If the percentage of optimization is greater than zero, output the adjustment result.
9. The grouped non-equal parallel machine scheduling method based on business rules according to claim 8, characterized in that, In step S4.17, determining whether the current solution is optimized compared to the original manually compiled solution specifically involves: These are manually compiled plan indicators. This is a scheme metric for a business rule-based grouped non-equal parallel machine scheduling method. The percentage of metrics optimized.
Citation Information
Patent Citations
Two-shift inverse multi-task scheduling optimization method considering different processing speeds
CN114186812A
Two-shift inverted multi-task scheduling optimization method considering switching time
CN114186813A
