MTO and MTS mixed production mode under the aps scheduling method
By combining the order expansion module, work order management module, and work order dispatch module with intelligent algorithms to optimize work order expansion and dispatch, the problem of unreasonable production planning and scheduling under the MTO-MTS hybrid production mode has been solved, and efficient production with rapid response to customer orders and inventory management has been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENYANG INST OF AUTOMATION - CHINESE ACAD OF SCI
- Filing Date
- 2021-12-17
- Publication Date
- 2026-05-12
AI Technical Summary
Existing technologies, when dealing with the MTO-MTS hybrid production mode, suffer from unreasonable production planning and scheduling, an inability to respond quickly to customer orders, resulting in inventory accumulation and low production efficiency.
The system employs an order expansion module, a work order management module, and a work order dispatch module, combined with intelligent algorithms to optimize the work order expansion and dispatch process. It considers order type, inventory material information, and equipment capacity constraints, and uses an immune genetic algorithm for work order dispatch. A mathematical model is established to minimize the maximum completion time and machine load.
It improved production efficiency, effectively solved the problem of inventory backlog, enabled rapid response to customer orders, and enhanced production management and economic benefits.
Smart Images
Figure CN116307008B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of management science and engineering, and involves the application of operations research modeling and intelligent optimization algorithms. Specifically, it is an APS scheduling method under a hybrid production mode of MTO and MTS. Background Technology
[0002] With the development of information technology, the production management level of most domestic enterprises has improved significantly. However, for enterprises with both MTO and MTS (Make-to-Stock) hybrid production modes, their production planning and scheduling remains relatively backward. Maintaining efficient continuous production in the face of MTS orders is a pressing issue. Meanwhile, for MTO (Make-to-Order) orders, enterprises cannot respond quickly to order demands, and cannot effectively address the problem of inventory accumulation. Production planning and scheduling are not entirely rational, failing to reconcile the contradiction between personalized user needs and rapid response, which greatly restricts the enterprise's production management level and economic efficiency.
[0003] For companies operating under a hybrid production model, the core of the scheduling process is how to break down different orders into work orders at different levels. This process requires considering order type, manufacturing MBOM information, and the status of semi-finished materials in inventory. How to quickly and effectively allocate these work orders to various machines while meeting both work order and equipment capacity constraints is also a crucial issue to consider during scheduling. Therefore, companies operating under a hybrid MTS-MTO production model need to consider the following factors during the scheduling process:
[0004] 1. How to quickly break down different types of orders into work orders at different levels, while taking into account both manufacturing MBOM and inventory material information.
[0005] 2. In the expanded work order set, how to define special work orders and how to determine the order of work order assignment.
[0006] 3. After the special work order is assigned, how to efficiently complete the work order assignment while meeting the delivery time requirements and equipment capacity, and thus complete the entire production scheduling process. Summary of the Invention
[0007] To address the shortcomings in production scheduling processes for enterprises with a hybrid MTS-MTO production model, this invention provides an ASP system scheduling method for enterprises exhibiting both inventory-based continuous batch production and market-oriented customer order-based discrete production characteristics, and develops a corresponding optimization system. The system consists of three main parts: an order expansion module, a work order management module, and a work order dispatch module. In the order expansion module, based on different order types and referring to MBOM and inventory material information, orders are expanded into work orders of different levels using different expansion methods. In the work order management module, the dispatch order of work orders is determined. In the work order dispatch stage, based on intelligent algorithms, constraints such as delivery date constraints, equipment capacity constraints, and work order correlation are satisfied to complete the dispatch process of all work orders, thereby completing the entire production scheduling.
[0008] The technical solution adopted by the present invention to achieve the above objectives is as follows:
[0009] An APS scheduling system for a hybrid production mode combining MTO and MTS includes:
[0010] The order expansion module is used to expand the input order collection into work orders of different levels based on the order category, manufacturing MBOM information and material inventory information, to obtain a work order set, and to determine the earliest start time and latest end time of each work order.
[0011] The work order management module is used to manage the production time of special work orders in the work order set and to sort the work orders.
[0012] The work order assignment module is used to assign each work order based on the sorting results of the work order management module and the set optimization goals and scheduling rules.
[0013] The special work order is a work order that can only be processed on specific equipment and at specific times.
[0014] An APS scheduling method for a hybrid production model combining MTO and MTS includes the following steps:
[0015] 1) The order expansion module expands the input order collection into work orders of different levels based on the order category, manufacturing MBOM information and material inventory information, obtains a work order set, and determines the earliest start time and latest end time of each work order;
[0016] 2) The work order management module manages the production time of special work orders in the work order set and sorts the work orders;
[0017] 3) The work order assignment module assigns each work order according to the sorting results of the work order management module and the set optimization goals and scheduling rules.
[0018] Step 1) specifically refers to:
[0019] If the order is a planned order, i.e. an order under the Make-to-Stock (MTS) model, the order will be expanded from the bottom layer of the Manufacturing Bill of Materials (MBOM) upwards, i.e. from raw materials to finished products. After the bottom layer work order is expanded, the next layer work order will be expanded until the top layer is reached.
[0020] If the order is a market-oriented order, i.e. a Make-to-Order (MTO) order, the order is expanded downwards from the top level of the Manufacturing Bill of Materials (MBOM), i.e. from finished products to raw materials. When the expanded work order cannot obtain raw materials from the inventory, it continues to expand downwards, associating the upper and lower work orders according to the execution sequence of the process, until the expanded work order can obtain raw materials from the inventory.
[0021] Step 2) includes the following steps:
[0022] 2.1) Modify work order attributes: Filter work orders in the work order set. If a work order is a special work order, lock the device that executes the work order and the time required for the device to execute the work order.
[0023] 2.2) Determine the order of work orders: special work orders are ordered first, and regular work orders are ordered according to the priority of their respective orders and the order of their respective processes in the process route.
[0024] Step 3) includes the following steps:
[0025] 3.1) Based on the set of work orders obtained in steps 1) and 2), as well as the processing equipment and processing sequence corresponding to each work order, construct a mathematical model for work order assignment;
[0026] 3.2) Based on the mathematical model of work order assignment, the work order assignment results are obtained using the immune genetic algorithm.
[0027] The mathematical model for work order assignment is as follows:
[0028] The production scheduling objective function is:
[0029] f1 = min(max) 1≤i≤n (C i )) (1)
[0030]
[0031] minf=ω1f1+ω2f2 (3)
[0032] The constraints are:
[0033] s ij +x ijk ×p ijk ≤c ij (4)
[0034] c ij ≤s i(j+1) (5)
[0035] c ij ≤C i (6)
[0036] s ij +p ijk ≤s i′j′ +L(1-y iji′j′k (7)
[0037] c ij ≤s i(j+1) +L(1-y i(j+1)i′j′k (8)
[0038]
[0039] Decision variables:
[0040] i=1, 2, ..., n; j=1, 2, ..., qj, k=1, 2, ..., m
[0041]
[0042]
[0043] Where f1: minimize the maximum completion time, f2: minimize the maximum machine load; f: weighted sum of the two optimization objectives; ω1 and ω2 are weighting coefficients; n: total number of orders; m: total number of machines; q: total number of work orders generated from n orders; C i : Completion time of order i; p ijk Work order p ij Processing time on machine k; ω1, ω2: weighting constants; i, i′: order number; k: machine number; j, j′: work order number; m ij : The number of selectable processing machines for the j-th work order derived from the i-th order; s ij Work order p ij Start processing time; c ij Work order p ij End processing time; L: coefficient;
[0044] Formula (1) is to minimize the maximum completion time; Formula (2) is to minimize the maximum machine load; Formula (3) is the optimization objective of the mathematical model;
[0045] Formulas (4) and (5) are the sequential constraints for each work order; Formula (6) is the work order completion time constraint; Formulas (7) and (8) are work order constraints, meaning that only one work order can be processed by the same machine at the same time; Formula (9) is the machine constraint, meaning that the same work order can only be processed by one machine at the same time.
[0046] The immune genetic algorithm includes the following steps:
[0047] a) Design of the coding method: The natural number segmentation coding method is adopted. Each code is composed of two parts, A and B, which represent the machine selection part and the work order sorting part of the work order, respectively. The value range of the machine selection part is 1, 2, ..., m, and the value range of the work order sorting part is 1, 2, ..., q. The chromosome length of both parts is equal to q, where q represents the total number of work orders.
[0048] b) Initial solution generation: All orders are processed according to their order type and work order expansion method to obtain a set of work orders including special work orders and ordinary work orders. Each work order can be selected to process a set of machines, and the work order is randomly matched to the corresponding machine to obtain the initial solution set of all work orders.
[0049] c) Crossover operation: Treat a group of A and B as an individual, and perform crossover operation on all individuals in a uniform manner, randomly selecting r gene loci for crossover at each position;
[0050] d) Mutation operation: All individuals are randomly mutated, and s positions are selected to mutate the machine selected by the work order at the corresponding position into any machine in its set of available machines;
[0051] e) Selection operation: Based on the optimization objective f of the work order assignment mathematical model, calculate the fitness value of each individual, and at the same time calculate the cross-entropy between each individual and the greedy optimal solution. Calculate the similarity between fitness value and cross-entropy, and select N individuals as the initial population for the next iteration process.
[0052] f) Repeat steps c) to e) until the iteration count requirement is met, and then obtain the optimal solution, which is the work order assignment result that minimizes the maximum completion time and the maximum machine load.
[0053] An APS scheduling system for a hybrid production mode of MTO and MTS includes a memory and a processor; the memory is used to store a computer program; the processor is used to implement the APS scheduling method for a hybrid production mode of MTO and MTS when the computer program is executed.
[0054] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the APS scheduling method under a hybrid MTO and MTS production mode.
[0055] The present invention has the following beneficial effects and advantages:
[0056] To address the production scheduling needs of enterprises with a hybrid production model encompassing both inventory-based continuous batch production and market-oriented discrete production based on customer orders, this paper introduces an advanced planning and scheduling optimization algorithm. The differences between market-oriented and planned orders are analyzed, and factors such as inventory information and product delivery dates are fully considered. Two different work order deployment methods are proposed to improve enterprise production efficiency. Constraints such as special orders and order priorities are also considered, with special work orders being prioritized in the work order management module. In the work order dispatch module, minimizing the maximum completion time and minimizing the maximum machine load are selected as optimization objectives. A mathematical model for the work order dispatch module is established and solved using an improved immune genetic algorithm. This allows for flexible selection of work order processing routes, effectively distributing each work order to each machine, comprehensively improving overall system performance, and completing the production scheduling plan. This has significant theoretical implications and broad application value for improving the flexibility of enterprise production management systems. Attached Figure Description
[0057] Figure 1 ASP production scheduling flowchart;
[0058] Figure 2 Flowchart of the order expansion module;
[0059] Figure 3 Work order management module flowchart;
[0060] Figure 4 Work order dispatch module flowchart;
[0061] Figure 5 Flowchart of the improved immune genetic algorithm for solving the work order assignment algorithm. Detailed Implementation
[0062] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.
[0063] The components of this system are described in detail below:
[0064] 1. Order Expansion Module
[0065] Using the combined set of all planned goods MTS orders and market goods MTO orders as input, and based on order category, manufacturing MBOM information, and inventory material information, the orders are broken down into work orders at different levels in different ways (e.g., Figure 2(as shown), and determine a rough production time range. The specific steps are as follows:
[0066] 1) Determine the order type. If it is a planned order, expand upwards from the bottom layer of the Manufacturing MBOM. After expanding the bottom-level work order, continue expanding the next work order until the top layer is reached. This expansion method ensures continuous production of the forecasted order.
[0067] 2) If the order is a market-oriented order, it is expanded downwards from the top level of the manufacturing MBOM. When the expanded work order cannot obtain raw materials from inventory, it continues to expand, linking the upper and lower work orders, until the expanded order can obtain raw materials from inventory. This expansion method avoids inventory backlog for market customers. When there is work-in-process in inventory, it can be used directly for production.
[0068] 2. Work Order Management Module
[0069] For the set of work orders obtained through the order expansion module, the work order management module processes certain special work orders and sorts them (e.g., ...). Figure 3 (As shown). The specific steps are as follows:
[0070] 2.1) Modify work order attributes. Filter work orders in the work order set. If a work order is a special work order, lock the corresponding equipment and the corresponding time for that work order.
[0071] 2.2) Determine the order of work orders. Special work orders are assigned first, and the remaining regular work orders are then sorted according to the priority of their respective orders and the order of their respective processes in the process route.
[0072] In actual production scheduling, situations are often quite complex, requiring users to manually modify some work order attributes to achieve semi-automatic scheduling. Therefore, in this module, users can filter and modify the attributes of some work orders generated by the order expansion module according to actual production needs. These are the special work orders, which can only be processed on specific equipment and at specific times.
[0073] 3. Work Order Dispatch Module
[0074] The main function of work order assignment is to distribute the work orders from the previous step to each piece of equipment according to their priority and in a manner that conforms to production scheduling constraints (e.g., ...). Figure 4 (As shown). Based on the order expansion module and work order management module, all work orders are assigned to various machines. Considering factors such as production process routes and machine load capacity, and with the optimization objectives of minimizing the maximum completion time and minimizing the maximum machine load, a mathematical model for the work order assignment module is established:
[0075] The production scheduling objective function is:
[0076] f1 = min(max) 1≤i≤n (C i )) (1)
[0077]
[0078] minf=ω1f1+ω2f2 (3)
[0079] The constraints are:
[0080] s ij +x ijk ×p ijk ≤c ij (4)
[0081] c ij ≤s i(j+1) (5)
[0082] c ij ≤C i (6)
[0083] s ij +P ijk ≤s i′j′ +L(1-y iji′j′k (7)
[0084] c ij ≤s i(j+1) +L(1-y i(j+1)i′j′k (8)
[0085]
[0086] Decision variables:
[0087] i=1, 2, ..., n; j=1, 2, ..., q j k = 1, 2, ..., m
[0088]
[0089]
[0090] Model parameters:
[0091] f1: Minimize the maximum completion time
[0092] f2: Minimize maximum machine load
[0093] f: a weighted sum of two optimization objectives, where ω1 and ω2 are weighting coefficients;
[0094] n: Total number of orders;
[0095] m: Total number of machines;
[0096] q: The total number of work orders generated from n orders;
[0097] C i : The completion time of order i;
[0098] p ijk Work order p ij Processing time on machine k
[0099] ω1, ω2: weighting constants;
[0100] i, i′: Order serial number;
[0101] k: Machine serial number;
[0102] j, j′: Work order number;
[0103] m ij The number of selectable processing machines for the j-th work order derived from the i-th order;
[0104] s ij Work order p ij Start processing time;
[0105] c ij Work order p ij End processing time;
[0106] L: A sufficiently large positive number;
[0107] Formula (1) is to minimize the maximum completion time; Formula (2) is to minimize the maximum machine load; Formula (3) is the optimization objective of the mathematical model.
[0108] Formulas (4) and (5) are the sequential constraints for each work order; Formula (6) is the work order completion time constraint; Formulas (7) and (8) are work order constraints, meaning that only one work order can be processed by the same machine at the same time; Formula (9) is the machine constraint, meaning that the same work order can only be processed by one machine at the same time.
[0109] The work order assignment problem in the above mathematical model belongs to the 0-1 integer programming problem. In practical applications, due to the large number and types of orders and work orders, and the large data scale, traditional mathematical programming methods are quite inefficient. Therefore, the improved immune genetic algorithm (IIGA) from intelligent algorithms is used to solve it. The steps are explained as follows:
[0110] 1. Design the coding method. Use a natural number segmented coding method. Each code consists of two parts, A and B, which represent the machine selection part and the work order sorting part of the work order, respectively. The value range of the machine selection part is 1, 2, ..., m, and the value range of the work order sorting part is 1, 2, ..., q. The chromosome length of both parts is equal to q, where q represents the total number of work orders.
[0111] 2. Initial solution generation: All orders are processed according to their type and work order expansion method, resulting in a work order pool containing special work orders and ordinary work orders. Each work order can be assigned to a set of machines for processing, and the order is randomly matched to the corresponding machine, thus obtaining the initial solution set for all work orders.
[0112] 3. Crossover operation: All individuals undergo a uniform crossover operation, with r gene loci randomly selected for crossover at each position to ensure that the new individual obtained after crossover is a feasible solution.
[0113] 4. Mutation operation: All individuals are randomly mutated. s positions are selected, and the machine selected by the work order at the corresponding position is mutated into any machine in its set of available machines.
[0114] 5. Select an operation: Based on the optimization objective f of the mathematical model of the work order assignment module, calculate the fitness value of each individual and the cross-entropy between each individual and the greedy optimal solution. Weigh the fitness value and the cross-entropy similarity, and select N individuals as the initial population for the next iteration.
[0115] 6. Repeat steps 3-5 until the iteration count requirement is met, and then obtain the optimal solution, which is the work order assignment result that minimizes the maximum completion time and the maximum machine load.
[0116] like Figure 1 As shown, this invention takes the actual production problems of a certain enterprise as the background, uses planned goods MTS orders and market goods MTO orders as input data, and combines an order expansion module and a work order management module to expand all orders to the equipment-level work order set according to different order types. Considering that the optimization objectives are to minimize the maximum completion time and minimize the maximum machine load, an intelligent scheduling algorithm for ASP system with mixed production characteristics of inventory-based continuous batch production and market customer order discrete production is designed and implemented. The specific implementation is described below:
[0117] Step 1: Input Data. The data consists of all order data for the company in the initial stage, including planned goods orders, market goods orders, and a few special orders (these are given directly as a work order pool and do not need to be expanded). Mark and categorize all orders for subsequent processing.
[0118] Step Two: Data Preprocessing. In this stage, the order set obtained in Step One is combined with the order expansion module and the work order management module to expand the orders into a set of work orders directly related to workshop production equipment. For planned goods orders, expansion proceeds upwards from the bottom layer of the Manufacturing MBOM. After expanding the bottom-level work orders, the next work order is expanded until the top layer is reached. For market goods orders, expansion proceeds downwards from the top layer of the Manufacturing MBOM. When an expanded work order cannot obtain raw materials from inventory, expansion continues, and the upper and lower work orders are linked until the expanded work order can obtain raw materials from inventory. For individual special orders, since they are given in the form of a work order collection, they can be directly decomposed into corresponding work orders. After expanding all orders into a work order pool, the work order management module is used to preprocess the work orders, including defining and assigning work order priorities, modifying the attributes of specific work orders, and locking the corresponding equipment.
[0119] Step 3: Input algorithm parameters. Specifically, these are the relevant parameters involved in the improved immune genetic algorithm (IIGA) of this invention, including the initial population size, number of algorithm iterations, crossover rate, mutation rate, etc., as well as the greedy optimal solution selection probability τ, which is improved in the local optimization part of this algorithm, and the weight parameters ω1 and ω2 of the two optimization objectives, etc.
[0120] Step 4: Utilize the improved immune genetic optimization algorithm described in this invention to generate an APS system scheduling plan for enterprises with mixed production characteristics, including both inventory-based continuous batch production and market-oriented customer order-based discrete production. This includes the following steps:
[0121] 1. Design coding method
[0122] The encoding method uses natural number segmentation. Each chromosome is divided into two parts, A and B, which represent the machine selection part and the work order sorting part of the work order, respectively. The length of both chromosome parts is equal to q, where q represents the total number of work orders.
[0123] X = [x 11 , ..., x 1j ;...;x n1 , ...x nj .x o11 , ..., x onj ]
[0124] The chromosome is divided into two parts. The first part is the machine selection part of the work order, where each small segment [x] i1 , ..., x ijThe first part represents the processing machine selected by each of the j work orders resulting from order i, and the second part is the work order sorting part, where the q work orders from the n orders are distributed across m machines. 011 , ..., x onj [This represents the order in which work orders are assigned; those listed earlier are assigned priority for processing.]
[0125] 2. Population initialization.
[0126] During the initialization phase, all input data is encoded according to the method described in section 1 to obtain the encoded set of the work order pool. Combined with the work order management module, specific work orders are processed specially, and equipment information and work order pool information are updated simultaneously to obtain a regular work order pool set R, forming an initial solution for subsequent optimization operations.
[0127] 3. Improved immune genetic algorithm for work order assignment
[0128] 3.1 Crossover Operation. Uniform crossover is used, with the following steps: 1) Randomly generate an integer r within the interval [1, T0]. 2) Generate r distinct integers based on the random number r, corresponding to the specific positions of the r invariant gene loci. 3) Using the integer r generated in step 2, copy the genes at the corresponding positions in parent chromosomes P1 and P2 to offspring chromosomes C1 and C2, maintaining their positions and order. 4) Copy the remaining genes from P1 and P2 to C2 and C1, maintaining their positions and order.
[0129] 3.2 Mutation Operation. The mutation part adopts a random mutation method, randomly selecting s positions in the mutated chromosome. For each position, the machine at each position is randomly replaced with one of the available machines from the set of available machines.
[0130] 3.3 Improve the immune genetic algorithm to enhance its overall performance
[0131] Define the greedy optimal solution X best This invention improves upon the immune genetic algorithm by introducing a greedy approach. Before each iteration of the algorithm, a greedy optimal solution X is found using this approach. best Calculate the relationship between the individual and X best The similarity is also discussed. The concept of cross-entropy is introduced. Cross-entropy is defined as:
[0132]
[0133] Where p(x) is the true probability distribution of the data, q(x) is the probability distribution calculated from the data, and cross-entropy is mainly used to measure the difference between two probability distributions.
[0134] The improved immune genetic algorithm in this invention obtains the greedy optimal solution X through a greedy approach in the initial stage of the population. best X best Corresponding to the true distribution p(x) in the cross-entropy concept, the other ordinary individuals in the population correspond to q(x). We expect individuals to be as similar as possible to X. best By calculating x for each individual i With X best The cross-entropy index is used to select the better individual during the selection process, thereby accelerating the optimization process of the entire algorithm.
[0135] Definitions of p(x) and q(x):
[0136]
[0137]
[0138] Define cross-entropy similarity:
[0139]
[0140] Cross-entropy similarity The larger the value, the higher the similarity between this solution and the greedy optimal solution.
[0141] 3.4 Selection Operation. Calculate the fitness values of the offspring individuals and their correlation with X. best Based on the cross-entropy value, select N individuals to proceed to the next iteration.
[0142] Repeat steps 3.1 to 3.4 until the set number of cycles is reached.
[0143] The flowchart of the entire production scheduling algorithm of the invention is as follows: Figure 5 As shown.
Claims
1. An APS scheduling method under a mixed production mode of MTO and MTS, characterized in that, Includes the following steps: 1) The order expansion module expands the input order collection into work orders of different levels based on the order category, manufacturing MBOM information and material inventory information, obtains a work order set, and determines the earliest start time and latest end time of each work order; 2) The work order management module manages the production time of special work orders in the work order set and sorts the work orders; 3) The work order assignment module assigns each work order according to the sorting results of the work order management module and the set optimization goals and scheduling rules; Step 1) specifically refers to: If the order is a planned order, i.e. an order under the Make-to-Stock (MTS) model, the order will be expanded from the bottom layer of the Manufacturing Bill of Materials (MBOM) upwards, i.e. from raw materials to finished products. After the bottom layer work order is expanded, the next layer work order will be expanded until the top layer is reached. If the order is a market order, i.e. a Make-to-Order (MTO) order, the order will be expanded from the top level of the Manufacturing Bill of Materials (MBOM) downwards, i.e. from finished products to raw materials. When the expanded work order cannot obtain raw materials from the inventory, it will continue to be expanded downwards, and the upper and lower work orders will be associated according to the execution order of the process, until the expanded work order can obtain raw materials from the inventory. Step 3) includes the following steps: 3.1) Based on the set of work orders obtained in steps 1) and 2), as well as the processing equipment and processing sequence corresponding to each work order, construct a mathematical model for work order assignment; 3.2) Based on the mathematical model of work order assignment, the work order assignment results are obtained using an immune genetic algorithm; The immune genetic algorithm includes the following steps: a) Coding method design: A natural number segmented coding method is adopted. Each code consists of two parts, A and B, representing the machine selection part and the work order sorting part, respectively. The value range of the machine selection part is... The range of values for the work order sorting section is: Both chromosome parts have a length equal to q, where q represents the total number of work orders; b) Initial solution generation: All orders are processed according to their order type and work order expansion method to obtain a set of work orders including special work orders and ordinary work orders. Each work order can be selected from a set of machines for processing, and the work order is randomly matched to the corresponding machine to obtain the initial solution set of all work orders. c) Crossover operation: Treat a group of A and B as an individual, and perform crossover operation on all individuals in a uniform manner, randomly selecting r gene loci for crossover at each position; d) Mutation operation: All individuals are randomly mutated, selecting s positions, and the machine selected by the work order at the corresponding position is mutated into any machine in its set of available machines; e) Selection operation: Based on the optimization objective f of the work order assignment mathematical model, calculate the fitness value of each individual, and at the same time calculate the cross-entropy between each individual and the greedy optimal solution. Calculate the similarity between fitness value and cross-entropy, and select N individuals as the initial population for the next iteration process. f) Repeat steps c) to e) until the iteration number requirement is met, and then obtain the optimal solution, which is the work order assignment result that satisfies the minimum maximum completion time and the minimum maximum machine load.
2. The APS scheduling method under a mixed production mode of MTO and MTS as described in claim 1, characterized in that, Step 2) includes the following steps: 2.1) Modify work order attributes: Filter work orders in the work order set. If a work order is a special work order, lock the device that executes the work order and the time required for the device to execute the work order. 2.2) Determine the order of work orders: special work orders are ordered first, and regular work orders are ordered according to the priority of their respective orders and the order of their respective processes in the process route.
3. The APS scheduling method under a mixed production mode of MTO and MTS as described in claim 1, characterized in that, The mathematical model for work order assignment is as follows: The production scheduling objective function is: （1） （2） （3） The constraints are: （4） （5） （6） （7） （8） （9） Decision variables: in, Minimize the maximum completion time. Minimize maximum machine load; : Weighted sum of two optimization objectives; n: Total number of orders; m: Total number of machines; q: Total number of work orders generated from n orders; : The completion time of order i; Work order Processing time on machine k; , Weighting constant; k: Order number; k: Machine serial number; Work order number; Work order Start processing time; Work order End processing time; L: coefficient; Formula (1) is to minimize the maximum completion time; Formula (2) is to minimize the maximum machine load; Formula (3) is the optimization objective of the mathematical model; Formulas (4) and (5) are the sequential constraints for each work order; Formula (6) is the work order completion time constraint; Formulas (7) and (8) are work order constraints, meaning that only one work order can be processed by the same machine at the same time; Formula (9) is the machine constraint, meaning that only one work order can be processed by one machine at the same time.
4. An APS scheduling system for a mixed production mode of MTO and MTS, used to implement the APS scheduling method for a mixed production mode of MTO and MTS as described in claim 1, characterized in that, include: The order expansion module is used to expand the input order collection into work orders of different levels based on the order category, manufacturing MBOM information and material inventory information, to obtain a work order set, and to determine the earliest start time and latest end time of each work order. The work order management module is used to manage the production time of special work orders in the work order set and to sort the work orders. The work order assignment module is used to assign each work order based on the sorting results of the work order management module and the set optimization goals and scheduling rules.
5. The APS scheduling system for a hybrid production mode of MTO and MTS as described in claim 4, characterized in that, The special work order is a work order that can only be processed on specific equipment and at specific times.
6. An APS scheduling system for a hybrid production mode of MTO and MTS, characterized in that, It includes a memory and a processor; the memory is used to store a computer program; the processor is used to implement, when executing the computer program, an APS scheduling method under a hybrid production mode of MTO and MTS as described in any one of claims 1-3.
7. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements an APS scheduling method under a hybrid production mode of MTO and MTS as described in any one of claims 1-3.