A scheduling method for continuous and discrete mixing processes
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SUPCON TECH CO LTD
- Filing Date
- 2022-12-30
- Publication Date
- 2026-07-03
Smart Images
Figure CN116205438B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of production scheduling technology, specifically to a scheduling method and system for a mixture of continuous and discrete processes. Background Technology
[0002] Existing large-scale production lines formulate production plans and schedules under the premise of limited resources. Based on the process route and bill of materials, they follow certain rules and optimization methods to find a feasible and optimal scheduling scheme. The feasibility of a scheduling scheme refers to meeting the quantity and time constraints of resources and materials. The goal of a scheduling scheme is usually a combination of multiple objectives, generally including balancing order delivery time, maximizing capacity, maximizing economic benefits, maintaining reasonable inventory levels of products and raw materials, and meeting customer priorities, among others.
[0003] The constraints and objectives of the production scheduling problem can be represented as a multi-objective mixed integer programming (MIP) problem consisting of multiple linear objective functions and a set of linear inequalities. The MIP problem can be gradually degenerated into an LP (Linear Programming) problem and solved by using branch and bound, genetic algorithms, or finite rules.
[0004] In large-scale fine chemical, daily chemical, and pharmaceutical production scenarios that simultaneously involve continuous and discrete production, continuous production lines produce large quantities of standardized basic products with multiple batch number switching capabilities. Discrete production lines, on the other hand, add and mix auxiliary materials and package products according to specific order requirements to produce thousands of SKU finished products. However, existing technologies do not address the connection between continuous and batch production lines in these scenarios, or they simply separate the continuous and batch production lines and schedule them as two separate parts. Furthermore, optimizing the scheduling algorithm still requires experience or manual guidance rather than global optimization. Summary of the Invention
[0005] In order to overcome the shortcomings of the above technologies, this invention provides a scheduling method and system for continuous and discrete mixed processes, which improves the scheduling connection constraint problem between continuous and discrete production lines, reduces the scale of scheduling problems, improves the scheduling optimization effect, and achieves the goal of improving production capacity efficiency.
[0006] The technical solution adopted by this invention to overcome its technical problems is as follows: This invention proposes a scheduling method for continuous and discrete mixed processes to acquire historical, static, and dynamic data of continuous and discrete production lines; determine the standard time granularity based on the acquired historical data, and calculate the batch consumption time of continuous and discrete production lines based on the standard time granularity; construct scheduling constraints based on static data, dynamic data, and batch consumption time, and construct a scheduling model with the scheduling objective function; solve the scheduling model based on the improved branch and bound algorithm and preset inventory verification constraints, thereby obtaining the production line scheduling plan for continuous and discrete mixed processes.
[0007] Furthermore, the step of determining the standard time granularity based on the acquired historical data and calculating the batch consumption time of continuous and discrete production lines based on the standard time granularity specifically includes: obtaining the average batch time Γ of each discrete production line based on historical data statistics. k The average shortest production cycle of a continuous production line Γ n And the original production batch time for each product on different production lines; based on the average batch time Γ k and the average shortest production cycle of continuous production lines Γ n Determine the standard time particle τ, where 1.25*min(Γ) k )≥τ≥0.75*min(Γ k And τ≥Γ n / a, a = 1, 2, 3, 4, 5; Calculate batch time β for continuous and discrete production lines based on standard time granularity τ. i，k ,in, β i，k This indicates the production batch time of product i on production line k. This indicates the time when product i was in the original production batch k on production line.
[0008] The above process completes the discretization of the continuous production line.
[0009] Furthermore, the production scheduling constraints based on static data, dynamic data, and batch time consumption specifically include: constructing inventory, production, and demand constraints based on product inventory and order information; constructing production line availability constraints based on the production line's availability status of products at time t in dynamic data; and constructing equipment conflict constraints based on the fact that a production line can only produce one type of product at the same time.
[0010] Furthermore, the production scheduling objective function is as shown in formula (1).
[0011] min objective=c1total_cap+c2stock_cost (1)
[0012] Wherein, min objective is the global objective value, total_cap is the maximum capacity as shown in formula (2), stock_cost is the minimum inventory cost as shown in formula (3), and C1 and C2 correspond to the global objective value and the objective weight for maximizing capacity, respectively.
[0013]
[0014]
[0015] Where, β i，k，t y represents the production batch time of product i after conversion on production line k at time t. i，k，t Indicates whether product i is produced on production line k at time t, with a value of 0 or 1, s i，t h represents the ending inventory of product i at time t. i，t This represents the inventory cost of product i at time t.
[0016] Furthermore, the preset inventory verification constraint s i，t This indicates that the ending inventory of product i at time t is within the upper and lower limits of the inventory, as shown in formula (4), where smin i，t smax i，t Indicates the upper and lower limits of inventory.
[0017] smin i，t ≤s i，t ≤smax i，t (4).
[0018] Establish a global constraint model that combines continuous and discrete production lines to achieve global constraint optimization of production scheduling.
[0019] Furthermore, the process of solving the production scheduling model based on the improved branch and bound algorithm and preset inventory verification constraints specifically includes: using inventory, production and demand constraints, production line availability constraints, and equipment conflict constraints as constraints of the production scheduling objective function; solving the production scheduling function based on the improved branch and bound algorithm; if a solution exists, determining whether the preset inventory verification constraints are satisfied; if satisfied, outputting the solution result; if not satisfied, using the unsatisfied preset inventory verification constraints as a constraint of the production scheduling model; if no solution exists, correcting the inventory, production, and demand constraints, and resolving the production scheduling function based on the improved branch and bound algorithm.
[0020] Furthermore, the production scheduling function is solved based on the improved branch and bound algorithm, specifically including: gradually tightening integer variables using the relaxation fixation method, and performing branch selection according to time order, node weight, and batch continuous time occupancy.
[0021] Furthermore, branch selection is performed based on batch continuous time occupancy processing, specifically including: production line k in a β i，k Only one batch of product i can be produced within the cycle, as shown in formula (5); after production line k plans product i at time t, β i，k No other products are produced during the cycle, as shown in formula (6); in variable y i，t，k When branching, for y i，t，k The branch = 1, while simultaneously setting y ii，tt，k ,ii∈F k , t≤tt≤t+β i，k Set to 0, where y i，t，k This indicates whether product i is produced on production line k at time t, and can take the value 0 or 1.
[0022]
[0023]
[0024] The branch and bound algorithm, which uses batch continuous time occupancy processing, improves the efficiency of solving the production scheduling problem.
[0025] The beneficial effects of this invention are:
[0026] 1. By continuously analyzing and extracting the features of the process route, the discretization strategy of the continuous production line is determined, thereby solving the problem of inconsistency in the time dimension between the continuous production line and the discrete production line in data modeling;
[0027] 2. Establish a global constraint model that combines continuous and discrete production lines to achieve global constraint optimization of production scheduling;
[0028] 3. By combining existing multi-objective fusion optimization strategies with branch-and-bound algorithms that handle continuous time occupancy of batches, the scheduling problem can be solved efficiently. Attached Figure Description
[0029] Figure 1 This is a flowchart of a scheduling method for a mixed continuous and discrete process according to an embodiment of the present invention;
[0030] Figure 2 This is a schematic diagram of the production line order information format according to an embodiment of the present invention;
[0031] Figure 3 Corresponding to the embodiments of the present invention Figure 2 A partial process route diagram for the order. Detailed Implementation
[0032] To facilitate a better understanding of the present invention by those skilled in the art, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The following are merely exemplary and do not limit the scope of protection of the present invention.
[0033] First, a list of definitions for some variables and symbols mentioned in this invention will be provided.
[0034] Materials: refers to the bill of materials, which represents some classification attributes of materials, such as finished products, semi-finished products, raw materials, etc.
[0035] BOM: A bill of materials used in the production of products.
[0036] To facilitate a better understanding of the present invention by those skilled in the art, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The following are merely exemplary and do not limit the scope of protection of the present invention.
[0037] like Figure 1 As shown in the flowchart, this embodiment describes a scheduling method for a mixed continuous and discrete process, which includes the following steps.
[0038] S1, acquire historical, static, and dynamic data from continuous and discrete production lines.
[0039] Static data includes at least process route information, BOM, material information, and equipment information. Dynamic data includes inventory information, start-up and shutdown information, and order information, while historical data includes at least historical order data and production data.
[0040] S2 determines the standard time granularity based on the acquired historical data, and calculates the batch consumption time for continuous and discrete production lines based on the standard time granularity.
[0041] S21 divides production lines into continuous production lines and discrete production lines according to their production characteristics.
[0042] In some implementations, a continuous production line is a production line that continuously feeds and discharges materials, while a discrete production line is a production line that feeds and discharges materials in batches.
[0043] S22, Calculate the average batch time Γ for each discrete production line. k and the average shortest production cycle of continuous production lines Γ n .
[0044] S23, based on average batch time Γ k and the average shortest production cycle Γ n The minimum time granularity τ of production scheduling is obtained, where 1.25*min(Γ) k )≥τ≥0.75*min(Γ k ), and τ≥Γ n / a, a = 1, 2, 3, 4, 5.
[0045] Furthermore, an integer time interval τ within a certain range is chosen as the granularity for the entire production scheduling. τ should not be too large, as this can easily lead to a loss of accuracy in small-batch production scheduling. Nor should it be too small, as this can result in excessively large problem sizes. Choosing an appropriate time granularity τ as the standard time granularity allows for a trade-off between model accuracy and computational efficiency.
[0046] S24, Calculate batch time β for continuous and discrete production lines based on standard time granularity τ. i，k ,in, β i，k This indicates the production batch time of product i on production line k. This indicates the time when product i was in the original production batch k on production line.
[0047] The above process completes the discretization of the production line.
[0048] S3 constructs scheduling constraints based on static data, dynamic data, and batch time consumption, and builds a scheduling model with the scheduling objective function.
[0049] S31, construct inventory, production, and demand constraints based on product inventory and order information.
[0050] The constraints on inventory, production, and demand are shown in Equation (1), which represents the balance of inventory, production, and demand for each material in each time interval.
[0051] s i，t-1 +x i，t =d i，t +s i，t (1)
[0052] Among them, s i，t s represents the inventory of product i at time t. i，0 Let x represent the initial inventory of product i. i，t Let d represent the planned production quantity of product i at time t, as shown in formula (2). i，t This represents the demand for product i at time t.
[0053]
[0054] in, y represents the production batch number of product i on production line k. i，t，k This indicates whether product i is produced on production line k at time t.
[0055] In some implementations, such as when there are multiple levels of processes and delivery times, the constraints of formula (1) can be expressed in the form of formula (3).
[0056]
[0057] Where, r i，j γ represents the quantity of product i required to produce one unit of product j. j This represents the time from the completion of production of product j to its delivery, where j∈D. i This indicates that product j is the direct raw material for product i.
[0058] S32, based on the availability of products at time t, construct the production line availability state constraints, as shown in formula (4).
[0059] y i，t，k ≤S t，k , i∈F k (4)
[0060] In equation (4), i∈F k This indicates that product i needs to be produced by production line k, s t，k This indicates the available status of production line k at time t.
[0061] S33, constructing equipment conflict constraints based on the premise that one production line can only produce one type of product at a time.
[0062] As shown in formula (5), this means that a production line can only produce one type of product at any given time.
[0063]
[0064] S34, within a complete production batch, the same equipment cannot produce the same product or other products, constraining the continuity of construction work.
[0065] This constraint mainly relates to the process of solving the objective function through branch and bound, starting from time zero and searching one standard time unit at a time, determining whether to produce or not as a branch. As shown in formulas (6) and (7), formula (6) indicates that production line k is in a β i，k Only one batch of product i can be produced within the cycle. Formula (7) indicates that after production line k plans product i at time t, β i，k No other products will be produced during the cycle.
[0066]
[0067]
[0068] S35, Construct the production scheduling objective function
[0069] As shown in equation (8), where min objective is the global objective value, total_cap is the maximum capacity as shown in equation (8), and stock_cost is the minimum inventory cost as shown in equation (9), C1 and C2 correspond to the global objective value and the objective weight for maximizing capacity, respectively.
[0070] min objective=c1total_cap+c2stock_cost (8)
[0071]
[0072]
[0073] Where, β i，k，t y represents the production batch time of product i after conversion on production line k at time t. i，k，t Indicates whether product i is produced on production line k at time t, with a value of 0 or 1, s i，t h represents the ending inventory of product i at time t. i，t This represents the inventory cost of product i at time t.
[0074] In an embodiment of the present invention, production scheduling constraints are defined and a production scheduling model is constructed with the production scheduling objective function, wherein the production scheduling constraints are used as constraints on the production scheduling function.
[0075] S4 solves the production scheduling model based on the improved branch and bound algorithm and preset inventory verification constraints, thereby obtaining the production line scheduling plan for continuous and discrete mixed processes.
[0076] Among them, the preset inventory verification constraint s i，t The ending inventory of product i at time t is within the upper and lower limits of the inventory, as shown in formula (11), where smin i，t smax i，t Indicates the upper and lower limits of inventory.
[0077] smin i，t ≤s i，t ≤smax i，t (11)
[0078] The flowchart illustrating the solution process for the production scheduling model based on the improved branch and bound algorithm and preset inventory verification constraints is as follows: Figure 1As shown, the production scheduling model is constructed as a MIP problem and solved. If a solution exists, it is determined whether the preset inventory verification constraint is met. If the preset inventory verification constraint is met, the solution result is output. If the preset inventory verification constraint is not met, the unmet preset inventory verification constraint is used as a constraint condition of the production scheduling model. If the MIP problem has no solution, the inventory, production, and demand constraints are modified, and the production scheduling function is solved again based on the improved branch and bound algorithm.
[0079] For the production scheduling plan proposed in this invention, the following strategy is designed for the improved branch delimitation method.
[0080] (1) Gradually tighten integer variables using the relaxation fixation method.
[0081] The characteristic of the production planning and scheduling problem is that production is scheduled in chronological order. The scheduling time period is divided into several intervals. According to the interval order, the integer variables of the previous intervals are fixed, only the integer variables in the current interval are optimized, the integer variables in the subsequent intervals are relaxed, and then the integer variables of the current interval are fixed based on the optimization results.
[0082] (2) Select branches according to priority. When selecting the next branch node, select according to time order and node weight in turn. This can ensure that the feasible solution with high weight is found first.
[0083] (3) Batch continuous time occupancy processing
[0084] There are two situations. The first situation is that after a production batch starts, the subsequent consecutive time periods will be occupied by that batch, as shown in formula (6). The second situation is that the same production line can only produce the same product at the same time, as shown in formula (7).
[0085] When performing a search, y is judged according to time, product priority, and production line priority. i，tt，k Whether this scalar is 0 or 1, once this variable is determined to be 1, then... The variable naturally becomes 0 due to constraint 1, which is the efficiency improvement achieved in the search strategy. Similarly, due to constraint 2... Other materials also have a value of 0 during this time period.
[0086] variable y i，t，k This variable indicates whether product i is being produced on production line k at time t. It only indicates the start time of production; in reality, a batch may span β. i，k Within a time interval, during which device k can no longer be scheduled for other production, due to the large number of constraints and their nonlinearity, this invention employs a heuristic integration approach. Specifically, regarding variable y... i，t，k When branching, for yi，t，k The branch = 1, while simultaneously setting y ii，tt，k ,ii∈F k , t≤tt≤t+β i，k Set to 0.
[0087] To better explain the production scheduling method for continuous and discrete mixed processes provided in the above embodiments of the present invention, a case study of organosilicon is used as an example. This example has been simplified to some extent. The order information format for this production line is as follows: Figure 2 As shown, the corresponding partial process route is as follows: Figure 3 As shown.
[0088] S1 first obtains basic data such as materials, BOM, and production line data, then obtains dynamic data such as inventory and order status, and finally obtains historical order data and production data.
[0089] The production line is divided into production lines and packaging lines, corresponding to the production of semi-finished products and finished products, respectively. There are many BOM levels, and there are many semi-finished products and substitute materials. Each route can be completed by approximately 3-10 production lines.
[0090] S2, will Figure 3 The production lines in the example are divided into continuous production lines and discrete production lines according to their production characteristics. In the example, the silicon powder, hydrolyzed material, production line, DMC, and VINYL production lines are continuous lines, while H48 and LSR are batch lines, i.e., discrete production lines.
[0091] Statistical analysis of the average batch time Γ for each discrete production line k and the average shortest production cycle of continuous production lines Γ n The minimum production cycle for a continuous production line is approximately 48 hours, while the minimum batch time for a batch production line is approximately 8 hours.
[0092] We choose 8 hours as the smallest time granularity τ for the entire production schedule.
[0093] Recalculate batch time β for continuous and discrete production lines using standard time τ. i，k Taking production line H48-1 as an example, the original production time was approximately 60 tons / 48 hours, which is now 60 tons / 6 hours. Production line body-1, the original production time was 10 tons / hour, which is now 60 tons / hour.
[0094] The above process completes the production line discretization process.
[0095] S3, Construct a production scheduling model
[0096] The constraints are constructed using inventory, production, and demand constraints as examples. In the case of multi-level processes and delivery time, formula (3) is used as the inventory, production, and demand constraints. Specifically, taking order 1 in the example as an example, for product LSR 9370, its subscript is ii, and for raw material H48V20000, its subscript is ij. The subscript of raw material H48V20000 is ik. The lead time γ of internally supplied raw materials is 3τ (24 hours). For each time period, products ii and ij have the constraints of formulas (12) and (13) as follows.
[0097]
[0098] This constraint requires that the production of product ii must be supported by the inventory of its raw material product ij.
[0099]
[0100] S4, the improved branch and bound method is used to solve the production scheduling model, thereby obtaining the production line scheduling plan for continuous and discrete mixed processes.
[0101] It should be noted that the steps of the corresponding methods are not necessarily performed in the order shown and described in this specification in other embodiments. In some other embodiments, the methods may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be broken down into multiple steps in other embodiments; and multiple steps described in this specification may be combined into a single step in other embodiments.
Claims
1. A scheduling method for continuous and discrete mixing processes, characterized in that, Specifically, it includes: Acquire historical, static, and dynamic data for both continuous and discrete production lines; The standard time granularity is determined based on the acquired historical data, and the batch time for continuous and discrete production lines is calculated based on the standard time granularity. Production scheduling constraints are constructed based on static data, dynamic data, and batch time consumption, and a production scheduling model is built with the production scheduling objective function. The production scheduling model is solved based on an improved branch and bound algorithm and preset inventory verification constraints, thereby... The production line scheduling plan for a mixture of continuous and discrete processes is obtained.
2. The scheduling method for a mixed continuous and discrete process according to claim 1, characterized in that, The process of determining the standard time granularity based on acquired historical data and calculating the batch time consumption for continuous and discrete production lines based on the standard time granularity specifically includes: The average batch time for each discrete production line was obtained based on historical data statistics. The average shortest production cycle of a continuous production line And the original production batch time for each product on different production lines; Based on average batch time and the average shortest production cycle of continuous production lines Determine the standard time granules ,in, and ; Based on standard time granules Calculate batch time for continuous and discrete production lines ,in, , This indicates the production batch time of product i on production line k. This indicates the time when product i was in the original production batch k on production line.
3. The scheduling method for a mixed continuous and discrete process according to claim 2, wherein... The characteristic is that, The production scheduling constraints constructed based on static data, dynamic data, and batch time consumption specifically include: Construct inventory, production, and demand constraints based on product inventory and order information; Construct constraints on the availability of products in the production line at time t based on dynamic data. Construct equipment conflict constraints based on the premise that a production line can only produce one type of product at a time.
4. A scheduling method for a mixed continuous and discrete process according to claim 3, characterized in that, The production scheduling objective function is shown in formula (1). (1) Where, min For the global target value, To maximize production capacity as shown in formula (2), To minimize inventory costs as shown in formula (3), C1 and C2 correspond to the global objective, respectively. Value and the target weight of maximizing capacity （2） （3） in, This represents the production batch time of product i after conversion on production line k at time t. This indicates whether product i is produced on production line k at time t, and takes the value 0 or 1. This represents the ending inventory of product i at time t. Indicates product In time Inventory costs.
5. A scheduling method for a mixed continuous and discrete process according to claim 4, characterized in that, The preset inventory verification constraint is: the ending inventory of product i at time t. The inventory must be within the upper and lower limits, as shown in formula (4), where, Indicates the upper and lower limits of inventory. (4)。 6. A scheduling method for a mixed continuous and discrete process according to claim 5, characterized in that, The solution to the production scheduling model based on the improved branch and bound algorithm and preset inventory verification constraints specifically includes: The constraints of inventory, production and demand, production line availability, and equipment conflict are used as the constraints of the production scheduling objective function. An improved branch and bound algorithm is used to solve the production scheduling function. If a solution exists, determine whether the preset inventory verification constraint is satisfied. If the preset inventory verification constraint is satisfied, output the solution result. If the preset inventory verification constraint is not satisfied, use the unsatisfied preset inventory verification constraint as a constraint of the production scheduling model. If no solution is found, the inventory, production, and demand constraints are modified, and the scheduling function is re-solved based on the improved branch and bound algorithm.
7. A scheduling method for a mixed continuous and discrete process according to claim 6, characterized in that, The production scheduling function is solved based on an improved branch and bound algorithm, which includes: gradually tightening integer variables using the relaxation fixation method, and selecting branches according to time order, node weight, and batch continuous time occupancy.
8. A scheduling method for a mixed continuous and discrete process according to claim 7, characterized in that, Branch selection based on batch consecutive time occupancy processing specifically includes: Production line k is in one Only one batch of product i can be produced within the cycle, as shown in formula (5); After production line k plans product i at time t, No other products are produced during the cycle, as shown in formula (6); In variables When branching, The branch, at the same time Set to 0, where, This indicates whether product i is produced on production line k at time t, and can take the value 0 or 1. （5） （6）。