Inventory-containing dynamic trademark production scheduling method
Through dynamic inventory consumption strategy and neighborhood search algorithm, the problems of inventory waste and idle equipment in chemical production are solved, and the efficient completion of production plans and high equipment utilization are achieved.
Patent Information
- Application Number
- CN202510793076.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-13
- Publication Date
- 2025-10-17
AI Technical Summary
Existing technologies have problems with inventory waste and idle processing equipment in chemical production. Especially in continuous production processes, the inventory priority consumption strategy may lead to excessive use of inventory for brands with low delay risks, and the processing equipment may be idle due to early completion.
A dynamic inventory consumption strategy is adopted. By setting the extension time to reduce the weight α, combined with the neighborhood search algorithm, the extension of the brand with high inventory selection index is prioritized, and the brand production plan is dynamically adjusted to balance inventory consumption and processing time.
It effectively avoids inventory waste, improves the utilization rate of processing equipment, and ensures the efficient completion of production plans.
Smart Images

Figure CN120806416A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of chemical production, in particular to grade production scheduling, and specifically to a dynamic grade production scheduling method with inventory. BACKGROUND
[0002] Polymer grade production scheduling is a core optimization problem in chemical production, especially in continuous production processes. Different grades of polymers (such as polyethylene, polypropylene, etc.) need to meet specific physical and chemical properties (such as melt index, density, molecular weight distribution, etc.) to adapt to downstream applications (such as film, injection molding, fiber, etc.). Each grade has an order delivery deadline, and a default penalty needs to be paid for exceeding the deadline, so the total overtime needs to be minimized. There is a certain amount of inventory for each grade, which can be used to buffer production fluctuations. The total processing time cannot exceed the processing cycle of the production plan, for example, if the processing cycle is 30 days, all production plans must be completed within 30 days.
[0003] Existing technical solutions
[0004] Inventory allocation: adopt the inventory-first consumption strategy, i.e., all grades use inventory first during processing, and the remaining demand enters the production queue. For example, if grade G1 requires 100 tons and the inventory is 30 tons, then the actual production of grade G1 is 70 tons;
[0005] Optimization goal: minimize the total overtime of all grades;
[0006] Hard constraint: the total processing time cannot exceed the processing cycle of the production plan;
[0007] Production scheduling algorithm: neighborhood search algorithm for grade production scheduling (existing solution).
[0008] Steps of the existing technical solution:
[0009] 1) Problem modeling
[0010] Objective function: minimize the total overtime of all grades
[0011] Hard constraint (cannot be violated): the total processing time cannot exceed the processing cycle of the production plan
[0012] 2) Inventory allocation strategy
[0013] All grades use inventory first during processing, and the remaining demand enters the production queue
[0014] 3) Initial solution generation
[0015] Randomly generate an initial production scheduling plan
[0016] 4) Neighborhood definition
[0017] Definition of "neighborhood" is a small adjustment of the current scheduling plan. The neighborhood is defined by the length of each grade overage, the less the length of each grade overage, the smaller the "neighborhood" of the grade, which will lead to better objective function; the more the length of each grade overage, the larger the "neighborhood" of the grade, which will lead to worse objective function.
[0018] 5) Neighborhood search
[0019] Find a better scheduling plan in the neighborhood of the current solution. If the neighborhood solution is better than the current solution, update the current solution.
[0020] 6) Termination condition
[0021] If the objective function value does not improve significantly within several iterations, or the solution time limit is reached, the algorithm terminates.
[0022] The above scheme has the following defects:
[0023] Inventory waste: forced to consume all inventory, which may lead to excessive use of inventory by grades with low delay risk, and no inventory to resist corresponding risks in case of sudden events in production and procurement.
[0024] Processing device idle: some grades are completed in advance due to inventory consumption, and the entire processing device completes the plan in advance, resulting in device idle.
[0025] For example: a device needs five kinds of grades G1-G5, and the processing days of each grade are shown in Table 1, a total of 33 days; the available inventory days are 11 days; after using all the inventory first, the actual processing days are accumulated to 22 days. The processing cycle of the device is 30 days, resulting in 8 days of idle time of the device.
[0026] Table 1 Example of existing scheme
[0027] Summary
[0028] The present application is directed to the problems in the background art, and proposes a dynamic grade scheduling method with inventory, which comprises the following steps:
[0029] 1) User sets the delay length reduction weight α, user inputs the grade scheduling basic information, and user inputs the inventory information;
[0030] 2) Calculate the inventory consumption time:
[0031] a. Total processing time of all grades represents the processing time of the i-th grade;
[0032] b. Total remaining inventory processing time represents the remaining inventory processing duration of the i-th grade;
[0033] c. consumed inventory duration Duration represents the processing cycle duration;
[0034] d. If P>0, i.e. the total processing duration exceeds the processing cycle duration, dynamic inventory consumption is required, go to 3); otherwise, use the neighborhood search algorithm to schedule the grades, go to 4);
[0035] 3) Co-cycle P rounds:
[0036] a. The current is the j-th round
[0037] b. Use the neighborhood search algorithm to schedule the grades;
[0038] c. Calculate the inventory balance index of all grades k represents the k-th grade of the j-th round; inventory balance index Obtained by:
[0039]
[0040] In the formula, represents the remaining inventory processing duration of the k-th grade of the j-th round;
[0041] d. Calculate the reduction amount Δp of the postponement duration of all grades j (G k );
[0042] e. Calculate the increase amount Δq of the advance duration of all grades j (G k );
[0043] f. Calculate the inventory consumption index s j (G k )=Δp j (G k )*α+Δq j (G k )*(1-α);
[0044] g. Calculate the inventory selection index
[0045] h. Select the k′-th grade with the maximum inventory selection index to consume one inventory unit;
[0046] i. Update the processing duration of all grades remaining inventory processing duration consumed inventory processing duration
[0047] j. Determine whether to proceed to P rounds, yes, go to 4), otherwise go to a;
[0048] 4) Output the brand production scheduling result, brand consumption inventory processing time;
[0049] 5) End.
[0050] Specifically, Δp j (G k ) and Δq j (G k ) are obtained by the following formula:
[0051]
[0052] Where, Δp j (G k ) represents the reduction of the postponement time of all brands after the jth round of production scheduling if brand G k consumes one unit of inventory time; Δq j (G k ) represents the increase in the advance time of all brands after the jth round of production scheduling if brand G k consumes one unit of inventory time; represents the processing completion time of the i th brand after the jth round of production scheduling; represents the difference between the processing completion time of the i th brand and the order delivery deadline after the jth round of production scheduling if brand G k consumes one unit of inventory time.
[0053] Specifically, the larger α is, the more inclined to reduce the postponement time of all brands in the entire production scheduling plan when consuming inventory; the smaller α is, the more inclined to increase the advance time of all brands in the entire production scheduling plan when consuming inventory.
[0054] Advantages of the present application
[0055] Dynamic inventory allocation: through multiple iterations, preferentially relieve the postponement of high inventory selection index brands, avoid inventory waste.
[0056] High utilization rate of processing device: avoid idle equipment due to advance completion of low priority tasks, improve overall utilization. BRIEF DESCRIPTION OF DRAWINGS
[0057] Figure 1 It is a brand production scheduling demand diagram in the embodiment.
[0058] Figure 2 It is a brand production scheduling result diagram without considering inventory consumption in the embodiment.
[0059] Figure 3The figure shows the production scheduling result of the first round of inventory consumption for the example. DETAILED DESCRIPTION
[0060] The application will be further described in connection with the examples, but the scope of the application is not limited thereto:
[0061] (1) Inventory consumption strategy
[0062] Each grade has a corresponding amount of inventory. To reduce the idle time of the device and avoid excessive consumption of inventory, only a part of the inventory of each grade needs to be consumed. First, there are countless possible combinations at this time. To reduce the number of combinations, all production is carried out in unit time, for example, 10 minutes can be used as unit time in actual production. Second, try to gradually consume the inventory, so as to gradually reduce the actual production time and gradually approach the processing cycle, while avoiding excessive consumption of inventory.
[0063] (2) Symbol definition
[0064] Duration: processing cycle time
[0065] G i : represents the i-th grade, a total of N grades, i = 1, 2,..., N;
[0066] represents the processing time required by the i-th grade in the j-th round, when j = 0, it represents the initial demand processing time of the grade;
[0067]
[0068] represents the remaining inventory processing time of the i-th grade in the j-th round, when j = 0, it represents the initial inventory processing time of the grade;
[0069]
[0070] represents the consumed inventory processing time of the i-th grade at the end of the j-th round, when j = 0, it represents the initial consumed inventory processing time of the grade, that is
[0071] Note: In actual production, the demand processing amount of the grade is converted into processing time through processing efficiency, and the unit of inventory is ton. For the sake of simplicity, all demand amounts are directly measured by "time" here.
[0072] t i : represents the order delivery deadline of the i-th grade;
[0073] represents the processing completion time of the i-th grade after the j-th round of production scheduling;
[0074] denotes the difference between the processing completion time and the order delivery deadline of the i-th brand after the i-th brand is arranged in the j-th round of production.
[0075]
[0076] When , it indicates that the brand is delayed in completion; when , it indicates that the brand is completed in advance.
[0077] (3) Inventory balance index
[0078] The existence of brand inventory can resist the corresponding risks of unexpected events in production and procurement, and can be used to adjust the actual processing length of the brand, so it is necessary to retain as much inventory as possible and to retain inventory as balanced as possible (i.e., some inventory is retained for each brand). The inventory balance index of the i-th brand in the j-th round is defined as
[0079]
[0080] That is, the proportion of the current remaining inventory processing length of a certain brand to the current remaining inventory processing length of all brands. The higher the inventory balance index of a certain brand, the more inventory that brand can consume relative to other brands, and the brand's inventory should be consumed first to ensure the balance of all brand inventories; the lower the inventory balance index of a certain brand, the less inventory that brand can consume relative to other brands, and the brand's inventory should be retained to consume the inventory of other brands as much as possible to ensure the balance of all brand inventories.
[0081] (4) Inventory consumption index
[0082] For a production plan in which the production of brands has been completed, if the total processing length at this time exceeds the processing cycle of the production plan, the inventory length needs to be consumed to shorten the total processing length. According to the "inventory consumption strategy", only one unit of inventory length is consumed at this time.
[0083] Because all brands have been arranged, it is noted that the consumption of one unit of inventory length by each brand affects the overall production plan. If brand G k consumes one unit of inventory length, the difference between the processing completion time and the order delivery deadline of other brands in production may change. For example, a certain brand G k+1 is produced after brand G k , and because brand G k consumes one unit of inventory length, brand G k will be completed in advance, and brand Gk+1 The processing completion time is also advanced accordingly.
[0084] represents the i-th brand after the j-th production scheduling, if the brand G k The difference between the processing completion time of the i-th brand and the order delivery deadline.
[0085] Δp j (G k ) represents the i-th brand after the j-th production scheduling, if the brand G k The reduction amount of the delay time of all brands when consuming one unit of inventory time.
[0086] Δq j (G k ) represents the i-th brand after the j-th production scheduling, if the brand G k The increase amount of the advance time of all brands when consuming one unit of inventory time.
[0087]
[0088] The customer determines the delay time reduction weight α (0 < α < 1), the larger α, the more inclined to reduce the delay time of all brands in the entire production scheduling when consuming inventory; the smaller α, the more inclined to increase the advance time of all brands in the entire production scheduling when consuming inventory.
[0089] Therefore, the i-th brand after the j-th production scheduling, if the brand G k Consume one unit of inventory time, the inventory consumption index of the new brand production scheduling is:
[0090] s j (G k ) = Δp j (G k )*α-Δq j (G k )*(1-α).
[0091] Finally, the application provides a dynamic brand production scheduling method with inventory, characterized in that it comprises the following steps:
[0092] 1) The user sets the delay time reduction weight α, inputs the brand production scheduling basic information, and inputs the inventory information;
[0093] 2) Calculate the inventory time that needs to be consumed:
[0094] a. The total processing time of all brands represents the processing time of the i-th brand;
[0095] b. The total remaining inventory processing time represents the remaining inventory processing duration of the i-th grade;
[0096] c. consumed inventory duration Duration represents the processing cycle duration;
[0097] d. If P>0, i.e. the total processing duration exceeds the processing cycle duration, dynamic inventory consumption is needed, go to 3); otherwise, use the neighborhood search algorithm to schedule the grades, go to 4);
[0098] 3) Co-cycle P rounds:
[0099] a. The current is the j-th round
[0100] b. Use the neighborhood search algorithm to schedule the grades;
[0101] c. Calculate the inventory balance index of all grades k represents the k-th grade in the j-th round; inventory balance index It is obtained by the following formula:
[0102]
[0103] In the formula, represents the remaining inventory processing duration of the k-th grade in the j-th round;
[0104] d. Calculate the reduction amount Δp of the postponement duration of all grades j (G k );
[0105] e. Calculate the increase amount Δq of the advance duration of all grades j (G k );
[0106] f. Calculate the inventory consumption index s j (G k )=Δp j (G k )*α+Δq j (G k )*(1-α);
[0107] g. Calculate the inventory selection index
[0108] h. Select the k′-th grade with the maximum inventory selection index to consume one inventory unit;
[0109] i. Update the processing duration of all grades remaining inventory processing duration consumed inventory processing duration
[0110] j. Determine whether to proceed to P rounds, yes, turn 4), otherwise turn a;
[0111] 4) Output the brand name production result, brand name consumption inventory processing time;
[0112] End.
[0113] The following will be combined with a specific embodiment, step 3) in the scheme is explained in detail:
[0114] A device needs five kinds of brand G1-G5, and the processing days of each brand are shown in Table 2, for example, the initial demand processing time of brand G1 is Remaining inventory processing time All brands need to process 33 days, and the inventory available days are 11 days. The processing cycle of the device is 30 days, which means at least 33-30=3 days of inventory length is consumed.
[0115] Table 2 Brand processing days
[0116]
[0117] Brand G1 needs to complete processing before No. 10, that is, t1=10; brand G2 needs to complete processing before No. 20, that is, t2=20; brand G3 needs to complete processing before No. 27, that is, t3=27; brand G4 needs to complete processing before No. 14, that is, t4=14; brand G5 needs to complete processing before No. 30, that is, t5=30. The brand production demand diagram is shown as Figure 1 .
[0118] When the inventory consumption is not considered, the brand production result is shown as Figure 2 , G1→G4→G2→G3→G5. Among them, brand G1 is completed at , which is 1 day later than the delivery deadline, that is Brand G4 is completed at , which is 3 days later than the delivery deadline, that is Brand G2 is completed at , which is 2 days later than the delivery deadline, that is Brand G3 is completed at , which is 1 day ahead of the delivery deadline, that is Brand G5 is completed at , which is 3 days later than the delivery deadline, that is
[0119] According to the production plan G1→G4→G2→G3→G5, the first round of all the grades consume one unit of inventory time (for example, 1 day as a unit). For example, the grade G1 consumes one unit of inventory time, so the completion time of the grade G1 is 1 day earlier, that is, just in the delivery period, so Similarly, the completion time of the grade G4 is 1 day earlier, so For the other three grades, in turn If the grade G2 consumes one unit of inventory time, since the grades G1 and G4 are in G2, the completion time is unchanged, so G2 and the grades after it are 1 day earlier, that is As shown in Table 3.
[0120] Table 3: Grades consume one unit of inventory time
[0121]
[0122] Further calculate the number of reduced delay time and the number of increased time of the grade, for example, the initial [G1, G4, G2, G3, G5] delay is [1, 3, 2, -1, 3]. When the grade G1 consumes one unit of inventory time, [G1, G4, G2, G3, G5] delay is [0, 2, 1, -2, 2], wherein the delay time of 4 grades (G1, G4, G2, G5) is reduced, that is, Δp 1 (G1) = 4; 1 grade (G3) is advanced, that is, Δq 1 (G1) = 1.
[0123] It is known that the customer determines the delay time reduction weight α = 0.8, which means that when consuming inventory, it is more inclined to reduce the number of delay time of all grades in the production plan.
[0124] Therefore, after the first round of production sorting, if the grade G1 consumes one unit of inventory time, the new inventory consumption index of the grade production plan is:
[0125] s 1 (G1) = Δp 1 (G1) * α - Δq 1 (G1) * (1-α) = 4 * 0.8 - 1 * 0.2 = 3
[0126] Similarly, the new inventory consumption index of the grade production plan can be calculated after the other grades consume one unit of inventory time, s 1 (G4) = 2.2, s 1 (G2) = 1.4, s 1 (G3) = 0.6, s 1(G5) = 0.8.
[0127] s j k ) = Δp j k ) * α + Δq j k ) * (1 - α)
[0128] (5) Inventory selection index
[0129] The inventory selection index of the kth brand after the jth round of production sequencing is constructed by combining the inventory balance index and the inventory consumption index:
[0130]
[0131] That is, the inventory balance index is greater, the inventory consumption index s j k is greater, the inventory selection index of a certain brand is greater, and the brand should consume a unit of inventory time in the current round; the inventory balance index is smaller, the inventory consumption index s j k is smaller, the inventory selection index of a certain brand is smaller, and the brand should not consume a unit of inventory time in the current round.
[0132] For example, in Table 4, if all brands consume a unit of inventory time in the first round, their respective inventory balance indexes, inventory consumption indexes, and calculated inventory selection indexes are given. The inventory consumption index of brand G1 has the highest value, but its inventory balance index is 0 (no available inventory), so its final inventory selection index is After comprehensive analysis, the inventory selection index of brand G2 is the highest, and a unit of inventory time should be consumed by this brand.
[0133] Table 4 Inventory selection index
[0134]
[0135] After the first round of inventory consumption, the processing days of the brands are shown in Table 5. The results of resequencing the production of the brands are shown in Figure 3 For example, the processing time required by brand G4 in the first round is the remaining inventory processing time the consumed inventory processing time at the end of the first round
[0136] Table 5 Processing days of brands after the first round of inventory consumption
[0137]
[0138] The specific embodiments described herein are merely illustrative of the spirit of the application. Various modifications or changes in addition or substitution to the described specific embodiments can be made by those skilled in the art without departing from the spirit of the application or exceeding the scope of the appended claims.
Claims
1. A dynamic brand production scheduling method with inventory, characterized in that It includes the following steps: 1) The user sets the delay time reduction weight α, the user enters the basic information of brand production scheduling, and the user enters the inventory information; 2) Calculate the time required to consume inventory: a. Total processing time for all grades Indicates the processing time of the i-th brand; b. Total remaining inventory processing time represents the processing time of the remaining inventory of brand i; c. Inventory consumption time Duration indicates the processing cycle length; d. If P>0, that is, the total processing time exceeds the processing cycle time, dynamic inventory consumption is required, and the 3); otherwise, use the neighborhood search algorithm to schedule brand production and go to 4); 3) Total cycle P rounds: a. The current round is j b. Use neighborhood search algorithm to schedule brand production; c. Calculate inventory balance indicators for all brands k represents the kth brand in the jth round; inventory balance indicator Obtained by the following formula: Where, represents the remaining inventory processing time of the kth brand in the jth round; d. Calculate the reduction in the extension time of all brands Δp j (G k ); e. Calculate the increase in the lead time of all brands Δq j (G k ); f. Calculate inventory consumption index s j (G k )=Δp j (G k )*α+Δq j (G k )*(1-α); g. Calculate inventory selection indicators h. Select the k′th brand with the largest inventory selection index and consume one inventory unit; i. Update the processing time of all grades Remaining inventory processing time Consume inventory processing time j. Determine whether to proceed to round P. If so, go to 4), otherwise go to a; 4) Output brand production scheduling results and brand consumption inventory processing time; 5) End.
2. The method according to claim 1, wherein: Δp j (G k ) and Δq j (G k ) is obtained by the following formula: Where Δp j (G k ) means that after the jth round of production sorting, if brand G k The amount of time it takes to consume one unit of inventory is the amount by which the extension time of all brands is reduced; Δq j (G k ) means that after the jth round of production sorting, if brand G k The amount of time it takes to consume one unit of inventory is the amount of time it takes to advance all brands; It represents the processing completion time of the brand i after the jth round of production sorting; Indicates that after the jth round of production sorting of brand i, if brand G k The time it takes to consume one unit of inventory is the difference between the processing completion time of the i-th brand and the order delivery deadline.
3. The method according to claim 1, wherein: The larger the α is, the more inclined it is to reduce the number of extension times for all brands in the entire production schedule when consuming inventory; the smaller the α is, the more inclined it is to increase the number of advance times for all brands in the entire production schedule when consuming inventory.