Order production scheduling management method based on MES system

Abstract

The invention relates to the technical field of digital MES systems, in particular to an MES system-based order production scheduling management method, which comprises the following steps of: detecting a production scheduling adjustment trigger signal, acquiring first production state data based on the MES system after the production scheduling adjustment trigger signal is acquired, and searching an optimal production scheduling strategy based on a preset genetic algorithm according to the first production state data. The method comprises the steps of obtaining a preliminary production scheduling plan, indicating a production line to produce based on the preliminary production scheduling plan, obtaining second production state data based on an MES system, and dynamically adjusting the preliminary production scheduling plan based on a preset reinforcement learning algorithm according to the second production state data to obtain a real-time production scheduling plan. And finally indicating the production line to produce based on the real-time production scheduling plan. According to the invention, the scheduling process is optimized cooperatively through the double intelligent algorithms, the global optimality of the production plan is ensured, the change of the production environment can be responded in real time, and the problem that the existing scheduling management method cannot respond quickly is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital MES systems, and in particular to an order scheduling management method based on an MES system. Background Art

[0002] In modern manufacturing, products typically undergo multiple complex processes and steps, from raw material input to finished product delivery. Irrational production planning and improper resource allocation can easily lead to problems such as poor process integration, equipment idling, or overloading. Furthermore, production sites are plagued by unexpected situations such as last-minute inserts, equipment failures, material shortages, and worker absences. These uncertainties further complicate production planning, making it difficult for traditional manual scheduling or simple automated scheduling methods to fully address the dynamic changes in production scenarios, leading to inefficient production processes and escalating costs.

[0003] Existing production scheduling management technologies primarily rely on ERP (Enterprise Resource Planning) systems or static Excel spreadsheets. Their core logic is to create fixed production plans based on preset rules and historical data. However, these approaches have significant limitations. When unexpected events (such as order changes, equipment anomalies, or supplier delays) occur, traditional systems are unable to detect the anomalies in real time and quickly adjust the plan. Manual intervention is required, and response times can often take hours or even days.

[0004] Therefore, people need an order scheduling management method based on the MES system to solve the problem that the existing scheduling management method cannot respond quickly. Summary of the Invention

[0005] The purpose of this invention is to provide an order scheduling management method based on the MES system to solve the following technical problems:

[0006] Existing production scheduling management methods cannot respond quickly.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] An order scheduling management method based on an MES system includes the following steps:

[0009] Detect production scheduling adjustment trigger signals;

[0010] After obtaining the production scheduling adjustment trigger signal, obtaining the first production status data based on the MES system;

[0011] According to the first production status data, an optimal production scheduling strategy is found based on a preset genetic algorithm to obtain a preliminary production scheduling plan;

[0012] Instruct the production line to start production based on the preliminary production schedule, and obtain the second production status data based on the MES system;

[0013] According to the second production status data, the preliminary production schedule is dynamically adjusted based on a preset reinforcement learning algorithm to obtain a real-time production schedule;

[0014] Instruct production lines to produce based on real-time production schedules.

[0015] As a further solution of the present invention: according to the first production status data, an optimal production scheduling strategy is found based on a preset genetic algorithm to obtain a preliminary production scheduling plan, including:

[0016] Creating a chromosome based on the first production status data, wherein the chromosome is used to represent the production schedule, and the chromosome includes a plurality of gene segments, each gene segment being used to represent the process allocation of an order;

[0017] Calculate the fitness of each chromosome based on the preset fitness function;

[0018] The chromosomes are screened, optimized and mutated according to fitness to obtain the optimal chromosome;

[0019] Based on the optimal chromosome, a preliminary production plan is obtained.

[0020] As a further solution of the present invention: based on a preset fitness function, the fitness of each chromosome is calculated, including:

[0021] Obtain target chromosome;

[0022] Calculate the total production duration of the production schedule represented by the target chromosome;

[0023] Calculate the equipment load variance of the production schedule represented by the target chromosome;

[0024] Calculate the order on-time rate of the production schedule represented by the target chromosome;

[0025] The fitness of the target chromosome is obtained by taking the weighted sum of the total production time, equipment load variance, and order on-time rate.

[0026] As a further solution of the present invention: according to the second production status data, the preliminary production schedule is dynamically adjusted based on a preset reinforcement learning algorithm to obtain a real-time production schedule, including:

[0027] establishing a state vector according to the second production state data;

[0028] Based on the preset action space and preset reward mechanism, reinforcement learning decision-making is performed according to the state vector to obtain a real-time production schedule;

[0029] The preset action space is a set of adjustment actions that can be performed on the production schedule, and the preset reward mechanism includes rewards for various results corresponding to the adjustment actions.

[0030] As a further solution of the present invention: the adjustment action in the preset action space includes at least one of maintaining the current production schedule, exchanging adjacent order processes, migrating orders to spare equipment, and delaying non-urgent orders.

[0031] As a further solution of the present invention: the preset reward mechanism includes positive rewards and negative penalties, wherein the positive rewards include rewards for completing one emergency order ahead of schedule and rewards for improving equipment utilization, and the negative penalties include penalties for equipment downtime and penalties for process conflicts.

[0032] As a further solution of the present invention: the first production status data includes static data and dynamic data, wherein the static data includes order data, equipment data, process data and material data, and the dynamic data includes real-time equipment status, production progress data and inventory dynamic change data; the second production status data includes equipment load rate, order priority vector, material inventory status and remaining process time forecast data.

[0033] As a further solution of the present invention: also include:

[0034] Obtaining a predicted delivery date based on the first production status data or the second production status data;

[0035] Send the forecast delivery date to the relevant departments.

[0036] The present invention also provides an order scheduling management system based on the MES system, comprising:

[0037] Status monitoring module, used to detect production scheduling adjustment trigger signals;

[0038] A data collection module is used to obtain first production status data based on the MES system after obtaining a production scheduling adjustment trigger signal;

[0039] A preliminary production scheduling module is used to find the optimal production scheduling strategy based on the first production status data and a preset genetic algorithm to obtain a preliminary production scheduling plan;

[0040] The production scheduling control module is used to instruct the production line to start production based on the preliminary production scheduling plan, and obtain the second production status data based on the MES system;

[0041] A real-time production scheduling module is used to dynamically adjust the preliminary production schedule based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production schedule;

[0042] The production scheduling control module is also used to instruct the production line to produce based on the real-time production scheduling plan.

[0043] The present invention also provides a computer-readable storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement any of the steps in the above-mentioned order scheduling management method based on the MES system.

[0044] Beneficial effects of the present invention:

[0045] The present invention provides an order production scheduling management method based on an MES system. The method first detects a production scheduling adjustment trigger signal. After obtaining the production scheduling adjustment trigger signal, the method obtains first production status data based on the MES system, searches for an optimal production scheduling strategy based on the first production status data and a preset genetic algorithm to obtain a preliminary production scheduling plan, and instructs the production line to produce based on the preliminary production scheduling plan. At the same time, the method obtains second production status data based on the MES system, dynamically adjusts the preliminary production scheduling plan based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production scheduling plan, and finally instructs the production line to produce based on the real-time production scheduling plan. The present invention collaboratively optimizes the production scheduling process through dual intelligent algorithms. The global search capability of the genetic algorithm quickly generates a preliminary production scheduling plan, effectively solving the problems of delayed response and low resource utilization of traditional production scheduling methods. The adaptive optimization characteristics of the reinforcement learning algorithm are used to continuously adjust the production scheduling plan, which significantly improves the flexibility and robustness of the system in responding to emergencies (such as equipment failures, order changes, etc.), and realizes closed-loop control from static scheduling to dynamic scheduling, which not only ensures the global optimality of the production plan, but also can respond to changes in the production environment in real time, thereby significantly shortening the order delivery cycle, reducing production costs and improving equipment utilization, solving the problem that existing production scheduling management methods cannot respond quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present invention will be further described below with reference to the accompanying drawings.

[0047] Figure 1 This is a method step diagram of the order scheduling management method based on the MES system of the present invention;

[0048] Figure 2 This is a system architecture diagram of the order scheduling management system based on the MES system of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0050] See also Figure 1As shown, the present invention is an order scheduling management method based on the MES system, comprising the following steps:

[0051] S101, detecting a production scheduling adjustment trigger signal;

[0052] S102: After obtaining a production scheduling adjustment trigger signal, obtaining first production status data based on the MES system;

[0053] S103. Finding an optimal production scheduling strategy based on the first production status data using a preset genetic algorithm to obtain a preliminary production scheduling plan;

[0054] S104, instructing the production line to start production based on the preliminary production schedule, and simultaneously obtaining second production status data based on the MES system;

[0055] S105. Dynamically adjust the preliminary production schedule based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production schedule.

[0056] S106. Instruct the production line to start production based on the real-time production schedule.

[0057] In the above process, the adjustment trigger signal refers to the signal that triggers the order scheduling adjustment, which can be any of the following situations:

[0058] (1) Sudden failure or maintenance of production equipment is triggered by the equipment status data (such as equipment shutdown, fault code, abnormal energy consumption, etc.) collected in real time by the MES system;

[0059] (2) Production progress deviation exceeds the limit, for example, the difference between the actual completion time of the current process and the planned time exceeds the preset threshold, which is detected by the MES system monitoring module and generates a signal;

[0060] (3) Changes in order priority, including urgent orders, customers’ temporary adjustments to delivery times, or order cancellations, are signaled by the ERP system or manual operations through the MES interface;

[0061] (4) Abnormal material supply, such as insufficient raw material inventory, delayed supplier delivery, or material quality issues, is triggered by the inventory management system or supplier collaboration platform;

[0062] (5) External environmental interference, such as fluctuations in energy supply, excessive temperature and humidity in the workshop, and other uncontrollable factors that affect production stability, are signaled by environmental monitoring sensors or energy management systems.

[0063] It is understandable that the above steps S104 and S106 can be repeatedly executed to continuously perform dynamic adjustments, and the real-time production scheduling plan currently being executed can be regarded as a preliminary production scheduling plan.

[0064] The present invention collaboratively optimizes the production scheduling process through dual intelligent algorithms. The global search capability of the genetic algorithm quickly generates a preliminary production scheduling plan, effectively solving the problems of delayed response and low resource utilization of traditional production scheduling methods. The adaptive optimization characteristics of the reinforcement learning algorithm are used to continuously adjust the production scheduling plan, which significantly improves the flexibility and robustness of the system in responding to emergencies (such as equipment failures, order changes, etc.), and realizes closed-loop control from static scheduling to dynamic scheduling, which not only ensures the global optimality of the production plan, but also can respond to changes in the production environment in real time, thereby significantly shortening the order delivery cycle, reducing production costs and improving equipment utilization, solving the problem that existing production scheduling management methods cannot respond quickly.

[0065] Furthermore, in a preferred embodiment, the above step S103, searching for an optimal production scheduling strategy based on a preset genetic algorithm according to the first production status data to obtain a preliminary production scheduling plan, specifically includes:

[0066] Creating a chromosome based on the first production status data, wherein the chromosome is used to represent the production schedule, and the chromosome includes a plurality of gene segments, each gene segment being used to represent the process allocation of an order;

[0067] Calculate the fitness of each chromosome based on the preset fitness function;

[0068] The chromosomes are screened, optimized and mutated according to fitness to obtain the optimal chromosome;

[0069] Based on the optimal chromosome, a preliminary production plan is obtained.

[0070] Specifically, in the above process, the first production status data includes static data and dynamic data, wherein the static data includes order data, equipment data, process data and material data, and the dynamic data includes real-time equipment status, production progress data and inventory dynamic change data.

[0071] Among them, order data includes data such as order ID, delivery date, priority and process route, which can be obtained through the ERP system. Equipment data includes data such as equipment ID, production capacity, available time and maintenance plan, which can be obtained through the equipment IoT sensor. Process data includes data such as process sequence and equipment-mold matching rules, which can be obtained through the PLM system. Material data includes data such as material coding, inventory level, consumption quota, etc., which can be obtained through the WMS system. The real-time status of the equipment includes data such as OEE, fault code, current load, etc., which can be obtained through IoT sensors. Production progress data includes data such as process completion rate and number of work-in-progress, which can be obtained through the MES work order system. Inventory dynamic change data includes data such as material in and out records and safety stock levels, which can be obtained through the WMS system.

[0072] Specifically, each chromosome consists of multiple gene segments, each of which corresponds to the process allocation of an order, for example: [Order A - Process 1 | Order A - Process 2 | Order B - Process 1 | ...]. Each gene segment can include the following fields: an order ID field, which uniquely identifies the order; a process ID field, which represents the process sequence number (e.g., SMT → CNC → assembly); an equipment ID field, which represents the assigned equipment number; a start time field, which represents the process start time (minute-level granularity); and a process parameter field, which represents key process parameters (temperature / pressure).

[0073] This embodiment deeply integrates the static data (such as order delivery date, equipment capacity, process route, material inventory) in the first production status data with the dynamic data (such as real-time equipment status, production progress, inventory changes) to construct a chromosome structure containing multi-dimensional information, so that the genetic algorithm can accurately represent the production scheduling plan under complex production scenarios. By dynamically integrating multi-source heterogeneous data such as ERP, PLM, WMS and IoT sensors, it solves the resource conflicts and efficiency bottlenecks caused by data fragmentation in traditional production scheduling methods. At the same time, it uses the global search capability of the genetic algorithm to effectively deal with the combination optimization problems under multiple constraints, and finally achieves high-precision generation and dynamic adaptive adjustment of production scheduling plans.

[0074] Furthermore, in a preferred embodiment, the above step of calculating the fitness of each chromosome based on a preset fitness function specifically includes:

[0075] Obtain target chromosome;

[0076] Calculate the total production duration of the production schedule represented by the target chromosome;

[0077] Calculate the equipment load variance of the production schedule represented by the target chromosome;

[0078] Calculate the order on-time rate of the production schedule represented by the target chromosome;

[0079] The fitness of the target chromosome is obtained by taking the weighted sum of the total production time, equipment load variance, and order on-time rate.

[0080] Specifically, the preset fitness function reflected in the above process can be expressed as follows:

[0081]

[0082] Here, F represents fitness, time represents total production time, load represents the variance of device load (i.e., the variance between the loads of individual devices), punctuality represents the order punctuality rate, and a, b, and c represent different parameter weights. For example, a can be set to 0.5 to prioritize delivery time, b can be set to 0.3 to prioritize load balancing, and c can be set to 0.2 to meet minimum quality compliance standards.

[0083] This embodiment significantly improves the comprehensive benefits and flexibility of production scheduling by designing a preset fitness function for multi-objective optimization. Specifically, it achieves a comprehensive evaluation of production efficiency, resource balance, and delivery reliability by quantifying the three core indicators of total production time, equipment load variance, and order punctuality, avoiding the problem of neglecting one thing while focusing on another due to single-indicator optimization. It adopts a weighted summation mechanism, allowing enterprises to dynamically adjust parameter weights according to actual needs (such as prioritizing delivery dates, focusing on load balancing, or meeting minimum compliance standards), so that production scheduling strategies can be flexibly adapted to different production scenarios and strategic goals. By incorporating equipment load variance into the optimization target, it effectively solves the problem of uneven equipment utilization in traditional methods, reduces the risk of local overload, and extends the service life of equipment.

[0084] Furthermore, in a preferred embodiment, the above step of dynamically adjusting the preliminary production schedule based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production schedule specifically includes:

[0085] establishing a state vector according to the second production state data;

[0086] Based on the preset action space and preset reward mechanism, reinforcement learning decision-making is performed according to the state vector to obtain a real-time production schedule;

[0087] The preset action space is a set of adjustment actions that can be performed on the production schedule, and the preset reward mechanism includes rewards for various results corresponding to the adjustment actions.

[0088] Specifically, in the above process, the second production status data includes equipment load rate, order priority vector, material inventory status and remaining process time forecast data, and the state vector is a vector composed of the above data.

[0089] In a preferred embodiment, the adjustment actions in the preset action space include at least one of maintaining the current production schedule, exchanging adjacent order processes, migrating orders to backup equipment, and delaying non-urgent orders.

[0090] For example, in the action space, the number 0 means maintaining the current production schedule, and the specific action can be no operation. The number 1 means exchanging adjacent order processes, and the specific action can be exchanging CNC equipment for order A and order B. The number 2 means migrating the order to the backup equipment, and the specific action can be migrating order A from injection molding machine #3 to #4. The number 3 means delaying non-urgent orders, and the specific action can be delaying the delivery date of order B by 2 hours.

[0091] In a preferred embodiment, the preset reward mechanism includes positive rewards and negative penalties, wherein positive rewards include rewards for completing one emergency order ahead of schedule and rewards for improving equipment utilization, and negative penalties include penalties for equipment downtime and penalties for process conflicts.

[0092] For example, in positive rewards, each urgent order completed ahead of schedule is worth +10 points, and a 5% increase in equipment utilization is worth +2 points. In negative penalties, equipment downtime exceeding 1 hour is worth -50 points, and process conflicts are worth -20 points.

[0093] This embodiment achieves dynamic adaptive optimization of production scheduling through a reinforcement learning algorithm, significantly improving the real-time responsiveness and anti-interference capabilities of the production system. Its core advantage lies in the construction of a closed-loop decision-making mechanism of state-action-reward. By quantifying key production status data such as equipment load rate, order priority, and material inventory into state vectors, it accurately depicts the dynamic production environment, enabling the algorithm to perceive production fluctuations in real time, pre-set an action space covering various adjustment actions (such as process exchange, equipment migration, order delays, etc.), and combine a refined reward mechanism of positive incentives (rewards for early completion of urgent orders, rewards for improving equipment utilization) with negative constraints (equipment downtime penalties, process conflict penalties), guiding the intelligent agent to quickly find the optimal adjustment strategy under complex constraints, effectively solving the problems of slow response to emergencies and single adjustment strategies of traditional production scheduling systems, and significantly improving the robustness and operational efficiency of the production system.

[0094] It is understandable that the specific process of the above genetic algorithm and reinforcement learning is an existing technology that can be understood by those skilled in the art, and therefore will not be described in detail in this article.

[0095] Furthermore, in a preferred embodiment, the order scheduling management method based on the MES system of the present invention further includes:

[0096] Obtaining a predicted delivery date based on the first production status data or the second production status data;

[0097] Send the forecast delivery date to the relevant departments.

[0098] Related departments are those whose work arrangements will be affected by adjustments to the production plan, such as sales and after-sales. For example, if a production machine suddenly breaks down, the MES recalculates the production schedule within 5 minutes, automatically migrates affected orders to backup equipment, and pushes the adjusted delivery date to the sales department.

[0099] This embodiment will actively push the predicted delivery date to related departments such as sales and after-sales, breaking the data silos of traditional production systems, allowing front-end departments to timely and synchronously adjust customer communication strategies (such as informing customers of delivery date changes in advance), avoiding customer complaints due to information lags, and achieving seamless connection between the production and business ends through the MES system, so that the sales department can make commitments to customers based on the latest delivery date, and the after-sales department can plan service resources in advance, forming a full-chain collaborative optimization from production planning to customer service, which not only improves the overall operational efficiency of the enterprise, but also enhances market competitiveness.

[0100] Combine Figure 2 As shown, the present invention also provides an order scheduling management system based on the MES system, comprising:

[0101] Status monitoring module 210, used to detect production scheduling adjustment trigger signal;

[0102] The data collection module 220 is configured to obtain first production status data based on the MES system after obtaining the production scheduling adjustment trigger signal;

[0103] A preliminary production scheduling module 230 is configured to find an optimal production scheduling strategy based on a preset genetic algorithm according to the first production status data to obtain a preliminary production scheduling plan;

[0104] The production scheduling control module 240 is used to instruct the production line to start production based on the preliminary production scheduling plan, and obtain the second production status data based on the MES system;

[0105] A real-time production scheduling module 250 is configured to dynamically adjust the preliminary production schedule based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production schedule;

[0106] The production scheduling control module 240 is also used to instruct the production line to perform production based on the real-time production scheduling plan.

[0107] It should be noted here that the corresponding system provided in the above embodiments is a computer program product, which can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiments, which will not be repeated here.

[0108] This embodiment further provides a computer-readable storage medium on which an order scheduling management program based on an MES system is stored. When the order scheduling management program based on the MES system is executed by a processor, the steps in the above embodiment can be implemented.

[0109] The present invention provides an order production scheduling management method based on an MES system. The method first detects a production scheduling adjustment trigger signal. After obtaining the production scheduling adjustment trigger signal, the method obtains first production status data based on the MES system, searches for an optimal production scheduling strategy based on the first production status data and a preset genetic algorithm to obtain a preliminary production scheduling plan, and instructs the production line to produce based on the preliminary production scheduling plan. At the same time, the method obtains second production status data based on the MES system, dynamically adjusts the preliminary production scheduling plan based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production scheduling plan, and finally instructs the production line to produce based on the real-time production scheduling plan. The present invention collaboratively optimizes the production scheduling process through dual intelligent algorithms. The global search capability of the genetic algorithm quickly generates a preliminary production scheduling plan, effectively solving the problems of delayed response and low resource utilization of traditional production scheduling methods. The adaptive optimization characteristics of the reinforcement learning algorithm are used to continuously adjust the production scheduling plan, which significantly improves the flexibility and robustness of the system in responding to emergencies (such as equipment failures, order changes, etc.), and realizes closed-loop control from static scheduling to dynamic scheduling, which not only ensures the global optimality of the production plan, but also can respond to changes in the production environment in real time, thereby significantly shortening the order delivery cycle, reducing production costs and improving equipment utilization, solving the problem that existing production scheduling management methods cannot respond quickly.

[0110] The above is a detailed description of an embodiment of the present invention. However, the content described is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An order scheduling management method based on MES system, characterized in that: The following steps are involved: Detect production scheduling adjustment trigger signals; After obtaining the production scheduling adjustment trigger signal, obtaining the first production status data based on the MES system; According to the first production status data, an optimal production scheduling strategy is found based on a preset genetic algorithm to obtain a preliminary production scheduling plan; Instruct the production line to start production based on the preliminary production schedule, and obtain the second production status data based on the MES system; According to the second production status data, the preliminary production schedule is dynamically adjusted based on a preset reinforcement learning algorithm to obtain a real-time production schedule; Instruct production lines to produce based on real-time production schedules.

2. The order scheduling management method based on the MES system according to claim 1 is characterized in that: According to the first production status data, the optimal production scheduling strategy is found based on the preset genetic algorithm to obtain a preliminary production schedule, including: Creating a chromosome based on the first production status data, wherein the chromosome is used to represent the production schedule, and the chromosome includes a plurality of gene segments, each gene segment being used to represent the process allocation of an order; Calculate the fitness of each chromosome based on the preset fitness function; The chromosomes are screened, optimized and mutated according to fitness to obtain the optimal chromosome; Based on the optimal chromosome, a preliminary production plan is obtained.

3. The order scheduling management method based on the MES system according to claim 2 is characterized in that: Based on the preset fitness function, the fitness of each chromosome is calculated, including: Obtain target chromosome; Calculate the total production duration of the production schedule represented by the target chromosome; Calculate the equipment load variance of the production schedule represented by the target chromosome; Calculate the order on-time rate of the production schedule represented by the target chromosome; The fitness of the target chromosome is obtained by taking the weighted sum of the total production time, equipment load variance, and order on-time rate.

4. The order scheduling management method based on the MES system according to claim 1 is characterized in that: According to the second production status data, the preliminary production schedule is dynamically adjusted based on the preset reinforcement learning algorithm to obtain a real-time production schedule, including: establishing a state vector according to the second production state data; Based on the preset action space and preset reward mechanism, reinforcement learning decision-making is performed according to the state vector to obtain a real-time production schedule; The preset action space is a set of adjustment actions that can be performed on the production schedule, and the preset reward mechanism includes rewards for various results corresponding to the adjustment actions.

5. The order scheduling management method based on the MES system according to claim 4 is characterized in that: The adjustment actions in the preset action space include at least one of maintaining the current production schedule, exchanging adjacent order processes, migrating orders to spare equipment, and delaying non-urgent orders.

6. The order scheduling management method based on the MES system according to claim 4 is characterized in that: The preset reward mechanism includes positive rewards and negative penalties. Positive rewards include rewards for completing each emergency order ahead of schedule and rewards for improved equipment utilization. Negative penalties include penalties for equipment downtime and penalties for process conflicts.

7. The order scheduling management method based on the MES system according to claim 1 is characterized in that: The first production status data includes static data and dynamic data, among which the static data includes order data, equipment data, process data and material data, and the dynamic data includes real-time equipment status, production progress data and inventory dynamic change data; the second production status data includes equipment load rate, order priority vector, material inventory status and remaining process time forecast data.

8. The order scheduling management method based on the MES system according to claim 1 is characterized in that: Also includes: Obtaining a predicted delivery date based on the first production status data or the second production status data; Send the forecast delivery date to the relevant departments.

9. An order scheduling management system based on MES system, characterized by: include: Status monitoring module, used to detect production scheduling adjustment trigger signals; A data collection module is used to obtain first production status data based on the MES system after obtaining a production scheduling adjustment trigger signal; A preliminary production scheduling module is used to find the optimal production scheduling strategy based on the first production status data and a preset genetic algorithm to obtain a preliminary production scheduling plan; The production scheduling control module is used to instruct the production line to start production based on the preliminary production scheduling plan, and obtain the second production status data based on the MES system; A real-time production scheduling module is used to dynamically adjust the preliminary production schedule based on the second production status data and a preset reinforcement learning algorithm to obtain a real-time production schedule; The production scheduling control module is also used to instruct the production line to produce based on the real-time production scheduling plan.

10. A computer-readable storage medium, characterized in that Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the order scheduling management method based on the MES system in any one of claims 1 to 8.

Images