Productivity balance type intelligent factory scheduling system based on MES (Manufacturing Execution System)

Abstract

The invention belongs to the technical field of factory production scheduling, and discloses an MES (Manufacturing Execution System)-based capacity balance type intelligent factory scheduling system, which comprises a preprocessing module for realizing the function of intelligently selecting a production strategy according to a current production mode, and improving the flexibility and accuracy of production decision making. The data execution module ensures the stable execution of the production plan, and can timely discover and respond to the change of the production plan, thereby avoiding the confusion and delay in the production process. And secondly, through introduction of the change scheduling module and the plan change module, the production plan change process is more scientific and efficient. By storing and calling a change algorithm, the system can automatically analyze the relation between order performance indexes and reasonably adjust the weight of each performance index of production, so that a new production plan better meeting actual requirements is formulated.

Description

Technical Field

[0001] The present invention belongs to the technical field of factory production scheduling, and relates to a capacity balancing intelligent factory scheduling system based on MES. Background Art

[0002] Electronic component manufacturers often accept orders for multiple varieties and small and medium-sized batches of electronic components. However, electronic component manufacturers often face the situation where production data between workshops in different factories is not communicated. In order to prevent shortage losses and inventory costs, how to reasonably formulate and arrange production plans according to order requirements and sales plans within the company's production capacity to achieve full utilization of production capacity, balanced production, and reduced waste is an urgent problem that electronic component manufacturers need to solve. Summary of the invention

[0003] The purpose of the present invention is to solve the problem in the prior art that electronic component manufacturers are unable to reasonably formulate and arrange production plans according to order requirements and sales plans, and to provide a capacity-balanced intelligent factory scheduling system and system based on MES.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions: The capacity-balanced intelligent factory scheduling system based on MES includes: pre-processing module, data execution module, change scheduling module, plan change module, cluster summary module, MES module, digital twin module and edge computing node module; The preprocessing module is used to select a production strategy according to the currently used production mode, and send the selected production strategy to the data execution module; The data execution module is used to store the currently used production plan algorithm and verify whether there is a production plan change. If so, the change scheduling module is called; if not, production is carried out according to the production strategy; The change scheduling module stores the designed production plan change algorithm and analyzes the relationship between various performance indicators of the order based on the pre-production scheduling; The plan change module changes the weights of various production performance indicators according to the adopted change algorithm, and calls the digital twin module to verify the scheduling feasibility, and then determines the new production plan and transmits it to the MES module; The MES module sends the new production plan to the data execution module for execution, monitors the workshop data, and uploads the collected workshop data to the cluster summary module; the cluster summary module collects the production plan and production progress of multiple workshops and multiple products in different locations of the enterprise, analyzes the historical scheduling data based on the current production history data, optimizes the weight distribution of each performance indicator and the priority of multiple targets, and controls the generation plan in real time; The edge computing node module is deployed in the workshop to collect equipment OEE, AGV path status, and quality inspection data in real time, and upload them to the MES module after local preprocessing.

[0005] A further improvement of the present invention is: Furthermore, local preprocessing is for the edge computing node module to remove invalid data from the collected equipment OEE, AGV path status and quality inspection data, and convert the collected equipment image binary stream into structured data; identify abnormal events through preset rules or lightweight models; and perform hashing or field masking on sensitive information; The workshop data monitored by the MES module include: the working status of production equipment, the current product processing quantity, product yield rate and scrap rate.

[0006] Furthermore, the various performance indicators include: delivery urgency, process complexity, resource occupancy intensity and economic benefit weight; The delivery urgency is the remaining buffer time divided by the standard processing cycle; The process complexity is the product of the process weight and the equipment accuracy requirement level; The resource occupancy intensity is the product of the result obtained by dividing the equipment occupancy time by the total generation cycle and the mold sharing conflict coefficient; The economic benefit weight is the product of the result obtained by dividing the order gross profit by the standard cost and the customer priority.

[0007] Furthermore, the relationship between the various performance indicators of the order is analyzed based on the pre-production scheduling. Specifically, the correlation between the various performance indicators is calculated based on the historical order data. According to the results of the correlation analysis, a dynamic correlation matrix reflecting the relationship between the performance indicators is constructed, specifically: Based on the Pearson correlation coefficient formula, the correlation coefficients between various performance indicators are obtained; Based on the correlation coefficients between various performance indicators, a dynamic correlation matrix reflecting the relationship between performance indicators is constructed; The Pearson correlation coefficient formula is: , in, and It is the first of two performance indicators. i Observations, and is their mean, n is the number of observations; Correlation coefficient r The value range is (-1, 1); when r When it is close to 1, it indicates that there is a strong positive correlation between the two performance indicators; whenr When it is close to -1, it indicates a strong negative correlation; when r When it is close to 0, it means there is almost no linear relationship; The construction of a dynamic correlation matrix reflecting the relationship between performance indicators is specifically as follows: If yes m performance indicators, then the correlation matrix is ​​a m × m A square matrix in which each element represents the correlation coefficient between two performance indicators; the elements on the diagonal are all 1.

[0008] Furthermore, the production plan change algorithm includes: shortest processing time principle, longest processing time principle, earliest due date principle, modified due date principle, first-come-first-served principle, last-come-first-served principle, minimum slack time and critical ratio principle; The shortest processing time principle first processes the job with the shortest processing time or expected processing time among all the processed jobs, focusing on giving priority to the task with the shortest processing time and maximizing the number of completed jobs; The longest processing time principle gives priority to the tasks with the longest processing time, which is suitable for production environments with generally long processing time and high requirements for overall process time optimization; The earliest due date principle is to ensure that the work with the key deadline is completed first, which is suitable for production environments with strict delivery requirements; The modified due date principle is to adapt to changes in production demand by adjusting the due date, which is suitable for production environments that require frequent adjustments to production plans and due dates; The first-come, first-served principle is that the jobs that arrive at the workshop first are processed first, which is suitable for production environments where the arrival time of jobs is relatively uniform and the waiting time is not very sensitive; The first-come, first-served principle means that the last-arrived job is processed first, which is suitable for urgent order processing; The minimum slack time is to give priority to jobs close to the deadline, which is suitable for production environments that are sensitive to job deadlines and can accurately predict the remaining working time and current time; The critical ratio principle ensures that critical tasks are given priority by comprehensively considering task allowances and remaining working time. It is suitable for production environments that have strict requirements on task priority and can accurately predict task allowances and remaining working time.

[0009] Furthermore, the change scheduling module also includes: emergency shutdown rules for serious defects in quality inspection and emergency insertion rules for high-priority orders. The emergency shutdown rules for serious defects in quality inspection are that when the quality inspection link finds that the products produced by the production line equipment have serious defects, the operation of the relevant production line equipment will be stopped immediately, and the production line equipment will be troubleshooted and repaired; if the repair is ineffective, it will be upgraded to a whole line shutdown, and the troubleshooting, repair and prevention mechanism will be started; the emergency insertion rules for high-priority orders are that for orders set to the highest priority, the current production plan will be adjusted, and the high-priority order will be inserted at the top of the production line; the production of emergency orders will be started, and the production progress and quality will be continuously monitored.

[0010] Furthermore, the weights of various production performance indicators are changed according to the adopted change algorithm, specifically: Dynamically calculate the weight of each indicator based on the entropy weight method-TOPSIS fusion model; Priority is sorted based on the weight of each indicator, and multi-objective optimization is performed through reinforcement learning algorithm to achieve the goal of minimizing entropy increase by changing the weight of each performance indicator.

[0011] Furthermore, the weights of each indicator are dynamically calculated based on the entropy weight method-TOPSIS fusion model, specifically: Based on the dynamic correlation matrix, obtain the positive index and negative index of the dynamic correlation matrix; Based on the obtained indicators, calculate the proportion of each sample under each indicator to obtain the proportion matrix; Based on the weight matrix, the entropy value and redundancy of each indicator are obtained, and the weight of each indicator is finally determined; Based on the standardized dynamic correlation matrix and weight vector, a weighted decision matrix is ​​obtained; Based on the weighted decision matrix, positive ideal solutions and negative ideal solutions are determined; The distance between each indicator and the positive ideal solution and the negative ideal solution is calculated to obtain the comprehensive score of each indicator. The indicators are sorted according to the comprehensive score to obtain the final sorting result.

[0012] Furthermore, the positive index of the dynamic correlation matrix is , The negative index of the dynamic correlation matrix is: , in, is the element in the i-th row and j-th column of the original data matrix, is the minimum value in the jth column, is the maximum value in the jth column; Calculate the proportion of each sample under each indicator, specifically: , in, is the sum of the standardized values ​​of all samples under this indicator; m is the number of rows and columns of the dynamic correlation matrix; The entropy value and redundancy of each indicator are obtained as follows: The entropy value of each indicator is: , in, , The redundancy of each indicator is , The weight of each indicator is determined by calculating the weight of each indicator according to the redundancy, specifically: , The weighted decision matrix is ​​obtained based on the standardized dynamic association matrix and weight vector, which is specifically: Given the normalized dynamic incidence matrix and the weight vector w =( w 1, w 2,…, wn ), weighted decision matrix V The calculation formula is: , The positive ideal solution and the negative ideal solution are determined based on the weighted decision matrix, specifically: The positive ideal solution is ; The positive index of the positive ideal solution is , The negative index of the positive ideal solution is , The negative ideal solution is ; The positive index of the negative ideal solution is , The negative index of the negative ideal solution is , The distance between each indicator and the positive ideal solution and the negative ideal solution is calculated to obtain the comprehensive score of each indicator, which is specifically: The distance between each indicator and the positive ideal solution for , The distance between each indicator and the negative ideal solution for , The calculation formula of the comprehensive score is: .

[0013] Furthermore, the performance indicators include: equipment utilization , Energy efficiency and changeover costs , and their corresponding weight coefficients α, β and γ; The weight of each performance indicator is changed by performing priority sorting based on the weight of each indicator and performing multi-objective optimization through a reinforcement learning algorithm, specifically: Set the state, action, reward and Q value update of the Q-learning algorithm; when Q-learning learns the optimal strategy, dynamically adjust the values ​​of α, β, and γ, and gradually achieve the required minimum entropy increase through iterative optimization; The state is the performance indicator value at the current moment ( , , ); the action is to adjust the values ​​of α, β, and γ; the Q value is updated by iteratively updating the Q value using the Q-learning update formula to find the optimal strategy, Among them, equipment utilization rate , Energy efficiency and changeover costs And the relationship between the corresponding weight coefficients α, β and γ is: , in, is the equipment utilization rate, For energy efficiency, is the conversion cost, weight coefficient , and Through Q-learning dynamic adjustment, the association matrix between the established indicators is used to select and optimize data, and the required entropy increase is minimized by adjusting the performance indicators. The reward function of the Q-learning algorithm introduces the digital twin simulation results. Specifically, if the capacity balance index CPBI of the scheduling plan after adjusting the weight increases by more than 10% in the simulated execution in the digital twin, the reward value increases; if the simulated execution leads to an increase in the quality risk factor, the reward value is reduced in a punitive manner; The reward function includes a capacity balance reward item and a quality risk penalty item, and is dynamically adjusted through weighting or conditional judgment; the expression is: , in, Positive rewards related to the improvement of the production capacity balance index CPBI; is the negative penalty associated with the increase in the quality risk factor; α and β are weight coefficients used to balance the priorities of production capacity and quality.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention realizes the function of intelligently selecting production strategies according to the current production mode through the preprocessing module, thereby improving the flexibility and accuracy of production decisions. The data execution module ensures the stable execution of the production plan, and can timely discover and respond to changes in the production plan, avoiding confusion and delays in the production process. Secondly, the introduction of the change scheduling module and the plan change module makes the production plan change process more scientific and efficient. By storing and calling the change algorithm, the system can automatically analyze the relationship between order performance indicators, reasonably adjust the weights of various production performance indicators, and thus formulate a new production plan that better meets actual needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 It is a flow chart of the capacity balancing intelligent factory scheduling system based on MES of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0018] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0020] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0021] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0022] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings: See also Figure 1 , the present invention discloses a capacity-balanced intelligent factory scheduling system based on MES, including: a preprocessing module, a data execution module, a change scheduling module, a plan change module, a cluster summary module, an MES module, a digital twin module and an edge computing node module; The preprocessing module is used to select a production strategy according to the currently used production mode, and send the selected production strategy to the data execution module; The data execution module is used to store the currently used production plan algorithm and verify whether there is a production plan change. If so, the change scheduling module is called; if not, production is carried out according to the production strategy; The change scheduling module stores the designed production plan change algorithm and analyzes the relationship between various performance indicators of the order based on the pre-production scheduling; The plan change module changes the weights of various production performance indicators according to the adopted change algorithm, and calls the digital twin module to verify the scheduling feasibility, and then determines the new production plan and transmits it to the MES module; The MES module sends the new production plan to the data execution module for execution, monitors the workshop data, and uploads the collected workshop data to the cluster summary module; the cluster summary module collects the production plan and production progress of multiple workshops and multiple products in different locations of the enterprise, analyzes the historical scheduling data based on the current production history data, optimizes the weight distribution of each performance indicator and the priority of multiple targets, and controls the generation plan in real time; The edge computing node module is deployed in the workshop to collect equipment OEE, AGV path status, and quality inspection data in real time, and upload them to the MES module after local preprocessing.

[0024] The digital twin module achieves closed-loop management of the entire life cycle of the production system through virtual-real mapping, simulation deduction and dynamic optimization. Its role is specifically reflected in the following multi-level functional architecture: (1) High-fidelity virtual model generation The digital twin module builds a full-factor virtual image covering equipment, production lines, and logistics paths based on the geometric structure (CAD model), physical properties (such as equipment thermodynamic parameters), and behavioral logic (such as AGV obstacle avoidance rules) of the physical factory. (2) Multi-source data fusion modeling Integrate the production planning data of the MES system, the equipment operation parameters of the SCADA system, and the order information of the ERP system to build a dynamic simulation environment driven by multi-dimensional data.

[0025] Use historical fault case libraries and real-time sensor data to train equipment degradation models to improve the prediction accuracy of virtual models MES receives equipment OEE data (equipment capability) (including green / yellow / red light duration, number of cycles, etc.) from edge nodes through the data acquisition module, and associates it with basic data such as equipment ledgers and process parameters to form a standardized equipment efficiency data set.

[0026] Local preprocessing is for the edge computing node module to remove invalid data from the collected equipment OEE, AGV path status and quality inspection data, and convert the collected equipment image binary stream into structured data; identify abnormal events through preset rules or lightweight models; hash or mask sensitive information; the workshop data monitored by the MES module include: the working status of production equipment, the current product processing quantity, product yield rate and scrap rate.

[0027] Among them, production strategy refers to the resource allocation rules and action plans formulated by enterprises under specific production modes to achieve goals such as cost, quality, delivery time, flexibility, etc. It can be simply divided into large-scale production, customized production, small-scale trial production, etc.

[0028] The production planning algorithm is a method that uses mathematical models or logical rules to generate the optimal production sequence under multiple constraints such as equipment capacity, material inventory, and delivery date. When a production plan change is detected, such as order adjustment or equipment failure, the production planning algorithm is changed to optimize the production strategy.

[0029] Equipment OEE is the overall efficiency of the equipment, which takes into account the three factors of equipment availability, performance and quality; availability refers to the proportion of time that the equipment is available in actual production, performance reflects the production rate of the equipment during the available time, and quality indicates the proportion of qualified products in the produced products.

[0030] The calculation formula of equipment OEE is: OEE = availability × performance × quality. Through this indicator, enterprises can clearly identify bottlenecks in the production process and make targeted optimizations.

[0031] AGV path status refers to the entire process of path planning, execution, and dynamic adjustment of the automated guided vehicle (AGV) when performing tasks. It is divided into fixed path guidance and free path guidance. Among them, fixed path guidance is to travel along a preset route (such as magnetic strips, light strips), which is suitable for structured environments. Free path guidance is to calculate the optimal path in real time through an algorithm to adapt to dynamic scenarios. Invalid values ​​of AGV path status include sensor failure, damaged navigation signs, and system failure, which result in the AGV being unable to accurately identify the path or obstacles, and may deviate from the predetermined route or collide.

[0032] The performance indicators include: delivery urgency, process complexity, resource occupancy intensity and economic benefit weight; The delivery urgency is the remaining buffer time divided by the standard processing cycle; The process complexity is the product of the process weight and the equipment accuracy requirement level; The resource occupancy intensity is the product of the result obtained by dividing the equipment occupancy time by the total generation cycle and the mold sharing conflict coefficient; The economic benefit weight is the product of the result obtained by dividing the order gross profit by the standard cost and the customer priority.

[0033] The relationship between the performance indicators of the order is analyzed based on the pre-production scheduling, specifically: based on the historical order data, the correlation between the performance indicators is calculated; according to the results of the correlation analysis, a dynamic correlation matrix reflecting the relationship between the performance indicators is constructed, specifically: Based on the Pearson correlation coefficient formula, the correlation coefficients between various performance indicators are obtained; Based on the correlation coefficients between various performance indicators, a dynamic correlation matrix reflecting the relationship between performance indicators is constructed; The Pearson correlation coefficient formula is: , in, and It is the first of two performance indicators. i Observations, and is their mean, n is the number of observations; Correlation coefficient r The value range is (-1, 1); when r When it is close to 1, it indicates that there is a strong positive correlation between the two performance indicators; when r When it is close to -1, it indicates a strong negative correlation; when r When it is close to 0, it means there is almost no linear relationship; The construction of a dynamic correlation matrix reflecting the relationship between performance indicators is specifically as follows: If yes m performance indicators, then the correlation matrix is ​​a m × m A square matrix in which each element represents the correlation coefficient between two performance indicators; the elements on the diagonal are all 1.

[0034] The production plan change algorithm includes: shortest processing time principle, longest processing time principle, earliest due date principle, modified due date principle, first-come-first-served principle, last-come-first-served principle, minimum slack time and critical ratio principle; The shortest processing time principle first processes the job with the shortest processing time or expected processing time among all the processed jobs, focusing on giving priority to the task with the shortest processing time and maximizing the number of completed jobs; The longest processing time principle gives priority to the tasks with the longest processing time, which is suitable for production environments with generally long processing time and high requirements for overall process time optimization; The earliest due date principle is to ensure that the work with the key deadline is completed first, which is suitable for production environments with strict delivery requirements; The modified due date principle is to adapt to changes in production demand by adjusting the due date, which is suitable for production environments that require frequent adjustments to production plans and due dates; The first-come, first-served principle is that the jobs that arrive at the workshop first are processed first, which is suitable for production environments where the arrival time of jobs is relatively uniform and the waiting time is not very sensitive; The first-come, first-served principle means that the last-arrived job is processed first, which is suitable for urgent order processing; The minimum slack time is to give priority to jobs close to the deadline, which is suitable for production environments that are sensitive to job deadlines and can accurately predict the remaining working time and current time; The critical ratio principle ensures that critical tasks are given priority by comprehensively considering task allowances and remaining working time. It is suitable for production environments that have strict requirements on task priority and can accurately predict task allowances and remaining working time.

[0035] The change scheduling module also includes: emergency shutdown rules for serious defects in quality inspection and emergency insertion rules for high-priority orders. The emergency shutdown rules for serious defects in quality inspection are that when the quality inspection link finds that the products produced by the production line equipment have serious defects, the operation of the relevant production line equipment is immediately stopped, and the production line equipment is troubleshooting and repair attempts are made; if the repair is ineffective, it is upgraded to a whole line shutdown, and the troubleshooting, repair and prevention mechanism is started; the emergency insertion rule for high-priority orders is that for orders set to the highest priority, the current production plan is adjusted, and the high-priority order is inserted at the top of the production line; the production of emergency orders is started, and the production progress and quality are continuously monitored.

[0036] The weight change of each production performance index according to the adopted change algorithm is specifically as follows: Dynamically calculate the weight of each indicator based on the entropy weight method-TOPSIS fusion model; Priority is sorted based on the weight of each indicator, and multi-objective optimization is performed through reinforcement learning algorithm to achieve the goal of minimizing entropy increase by changing the weight of each performance indicator.

[0037] The entropy weight method-TOPSIS fusion model dynamically calculates the weight of each indicator, specifically: Based on the dynamic correlation matrix, obtain the positive index and negative index of the dynamic correlation matrix; Based on the obtained indicators, calculate the proportion of each sample under each indicator to obtain the proportion matrix; Based on the weight matrix, the entropy value and redundancy of each indicator are obtained, and the weight of each indicator is finally determined; Based on the standardized dynamic correlation matrix and weight vector, a weighted decision matrix is ​​obtained; Based on the weighted decision matrix, positive ideal solutions and negative ideal solutions are determined; The distance between each indicator and the positive ideal solution and the negative ideal solution is calculated to obtain the comprehensive score of each indicator. The indicators are sorted according to the comprehensive score to obtain the final sorting result.

[0038] The positive index of the dynamic correlation matrix is , The negative index of the dynamic correlation matrix is: , in, is the element in the i-th row and j-th column of the original data matrix, is the minimum value in the jth column, is the maximum value in the jth column; Calculate the proportion of each sample under each indicator, specifically: , in, is the sum of the standardized values ​​of all samples under this indicator; m is the number of rows and columns of the dynamic correlation matrix; The entropy value and redundancy of each indicator are obtained as follows: The entropy value of each indicator is: , in, , The redundancy of each indicator is , The weight of each indicator is determined by calculating the weight of each indicator according to the redundancy, specifically: , The weighted decision matrix is ​​obtained based on the standardized dynamic association matrix and weight vector, which is specifically: Given the normalized dynamic incidence matrix and the weight vector w =( w 1, w 2,…, wn ), weighted decision matrix V The calculation formula is: , The positive ideal solution and the negative ideal solution are determined based on the weighted decision matrix, specifically: The positive ideal solution is ; The positive index of the positive ideal solution is , The negative index of the positive ideal solution is , The negative ideal solution is ; The positive index of the negative ideal solution is , The negative index of the negative ideal solution is , The distance between each indicator and the positive ideal solution and the negative ideal solution is calculated to obtain the comprehensive score of each indicator, which is specifically: The distance between each indicator and the positive ideal solution for , The distance between each indicator and the negative ideal solution for , The calculation formula of the comprehensive score is: .

[0039] The performance indicators include: equipment utilization , Energy efficiency and changeover costs , and their corresponding weight coefficients α, β and γ; The weight of each performance indicator is changed by performing priority sorting based on the weight of each indicator and performing multi-objective optimization through a reinforcement learning algorithm, specifically: Set the state, action, reward and Q value update of the Q-learning algorithm; when Q-learning learns the optimal strategy, dynamically adjust the values ​​of α, β, and γ, and gradually achieve the required minimum entropy increase through iterative optimization; The state is the performance indicator value at the current moment ( , , ); the action is to adjust the values ​​of α, β, and γ; the Q value is updated by iteratively updating the Q value using the Q-learning update formula to find the optimal strategy, Among them, equipment utilization rate , Energy efficiency and changeover costs And the relationship between the corresponding weight coefficients α, β and γ is: , in, is the equipment utilization rate, For energy efficiency, is the conversion cost, weight coefficient , and Through Q-learning dynamic adjustment, the association matrix between the established indicators is used to select and optimize data, and the required entropy increase is minimized by adjusting the performance indicators. The reward function of the Q-learning algorithm introduces the digital twin simulation results. Specifically, if the capacity balance index CPBI of the scheduling plan after adjusting the weight increases by more than 10% in the simulated execution in the digital twin, the reward value increases; if the simulated execution leads to an increase in the quality risk factor, the reward value is reduced in a punitive manner; The reward function includes a capacity balance reward item and a quality risk penalty item, and is dynamically adjusted through weighting or conditional judgment; the expression is: , in, Positive rewards related to the improvement of the production capacity balance index CPBI; is the negative penalty associated with the increase in the quality risk factor; α and β are weight coefficients used to balance the priorities of production capacity and quality.

[0040] Example: The present invention discloses a capacity-balanced intelligent factory scheduling method based on MES, which acts on a capacity-balanced intelligent factory scheduling system based on MES, and includes: According to the sales order input preprocessing module, production is carried out according to the currently set algorithm strategy. At the same time, the equipment, product quantity, processing process, estimated completion time, and warehousing, outbound and delivery time estimated to be occupied by this order number are reported, and then sent to the data execution module for verification. When there is no change in the production plan after the data execution module has verified it, it is sent to the MES module for production according to the order production plan.

[0041] When the data execution module finds that the order needs to be adjusted and the production plan changes, the production algorithm is selected through manual decision-making, and the data is transmitted to the change scheduling module. The change scheduling module performs pre-production analysis and preliminarily estimates the changes in all currently scheduled production plans after the production plan is changed. If there is no manual objection, the data is transmitted to the plan change module after the order is selected, and the weight analysis of various indicators is performed to complete the capacity balance adjustment, and the digital twin module is called to verify the feasibility of scheduling, and the data is transmitted to the MES module for production execution. If there is a manual objection, the order performance indicators can be adjusted in part after the pre-production scheduling, and the order selection can be performed after the adjustment, so that the data can be transmitted to the plan change module.

[0042] The edge computing node module is deployed in the workshop to collect equipment OEE, AGV path status, and quality inspection data in real time, and upload it to the MES module after local preprocessing.

[0043] If the quality inspection failure rate of a production line exceeds the warning level or the product production quantity exceeds the warning level, it will be treated as a production failure. This part of the data will be reported by the MES system. Once a failure is reported, the scheduling module will be changed to temporarily lower the weight of the faulty production line. Once the manual processing is completed and the production line returns to normal, the scheduling module will be changed to restore the weight adjustment of each production line before the failure.

[0044] After receiving the production plan change, the change scheduling module needs to further formulate the weights of various indicators of the production line, and estimate the change time of the current production order after the adjustment of the production line. After no manual objection, it will be directly transmitted to the MES module for execution. If there is a manual objection, the weight of the production line indicator can be adjusted separately to adjust the order production situation.

[0045] The production situation of multiple products and multiple workshops in different locations of an enterprise is collected through the MES module and then aggregated to the cluster summary module, which can display the production progress and production plan of all the multiple workshops in different locations of the entire enterprise. The cluster summary module can also make changes to the data execution module and the change scheduling module, and can, for example, match the source of goods and the production address to ensure the lowest logistics cost and the highest production efficiency.

[0046] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The capacity-balanced intelligent factory scheduling system based on MES is characterized by: include: Preprocessing module, data execution module, change scheduling module, plan change module, cluster summary module, MES module, digital twin module and edge computing node module; The preprocessing module is used to select a production strategy according to the currently used production mode, and send the selected production strategy to the data execution module; The data execution module is used to store the currently used production plan algorithm and verify whether there is a production plan change. If so, the change scheduling module is called; If it does not exist, production is performed according to the production strategy; The change scheduling module stores the designed production plan change algorithm and analyzes the relationship between various performance indicators of the order based on the pre-production scheduling; The plan change module changes the weights of various production performance indicators according to the adopted change algorithm, and calls the digital twin module to verify the feasibility of scheduling, and then determines the new production plan and transmits it to the MES module; The MES module sends the new production plan to the data execution module for execution, monitors the workshop data, and uploads the collected workshop data to the cluster summary module; The cluster summary module collects the production plans and production progress of multiple workshops and multiple products in different locations of the enterprise, analyzes the historical scheduling data based on the current production history data, optimizes the weight distribution of each performance indicator and the priority of multiple objectives, and controls the generation plan in real time; The edge computing node module is deployed in the workshop to collect equipment OEE, AGV path status, and quality inspection data in real time, and upload them to the MES module after local preprocessing.

2. The capacity balancing intelligent factory scheduling system based on MES according to claim 1 is characterized in that: The local preprocessing is that the edge computing node module removes invalid data from the collected equipment OEE, AGV path status and quality inspection data, and converts the collected equipment image binary stream into structured data; Identify abnormal events through preset rules or lightweight models; Hash or mask sensitive information; The workshop data monitored by the MES module include: the working status of production equipment, the current product processing quantity, product yield rate and scrap rate.

3. The capacity balancing intelligent factory scheduling system based on MES according to claim 2 is characterized in that: The performance indicators include: delivery urgency, process complexity, resource occupancy intensity and economic benefit weight; The delivery urgency is the remaining buffer time divided by the standard processing cycle; The process complexity is the product of the process weight and the equipment accuracy requirement level; The resource occupancy intensity is the product of the result obtained by dividing the equipment occupancy time by the total generation cycle and the mold sharing conflict coefficient; The economic benefit weight is the product of the result obtained by dividing the order gross profit by the standard cost and the customer priority.

4. The capacity balancing intelligent factory scheduling system based on MES according to claim 3 is characterized in that: The relationship between the performance indicators of the order according to the pre-production analysis is specifically: based on the historical order data, the correlation between the performance indicators is calculated; according to the results of the correlation analysis, a dynamic correlation matrix reflecting the relationship between the performance indicators is constructed, specifically: Based on the Pearson correlation coefficient formula, the correlation coefficients between various performance indicators are obtained; Based on the correlation coefficients between various performance indicators, a dynamic correlation matrix reflecting the relationship between performance indicators is constructed; Among them, the correlation coefficient r The value range is (-1, 1); when r When it is 1, it means that the strong positive correlation between the two performance indicators is the largest; r Cannot be 1, so r When the value of approaches 1 infinitely, the stronger the positive correlation between the two performance indicators is; r As the value of r When it is 0, it means there is no relationship between the two performance indicators. r When it is -1, it means that the two performance indicators have a strong negative correlation. Similarly, due to r It cannot be -1, so r When the value of is infinitely close to -1, the stronger the negative correlation between the two performance indicators is; The construction of a dynamic correlation matrix reflecting the relationship between performance indicators is specifically as follows: If yes m performance indicators, then the correlation matrix is ​​a m × m A square matrix in which each element represents the correlation coefficient between two performance indicators; the elements on the diagonal are all 1.

5. The capacity balancing intelligent factory scheduling system based on MES according to claim 4 is characterized in that: The production plan change algorithm includes: shortest processing time principle, longest processing time principle, earliest due date principle, modified due date principle, first-come-first-served principle, last-come-first-served principle, minimum slack time and critical ratio principle; The shortest processing time principle first processes the job with the shortest processing time or expected processing time among all the processed jobs, focusing on giving priority to the task with the shortest processing time and maximizing the number of completed jobs; The longest processing time principle gives priority to the tasks with the longest processing time; The earliest due date principle is to ensure that the work with the key deadline is completed first, which is suitable for production environments with strict delivery requirements; The modified due date principle is to adapt to changes in production demand by adjusting the due date, which is suitable for production environments that require frequent adjustments to production plans and due dates; The first-come, first-served principle is that the jobs that arrive at the workshop first are processed first; The first-come, first-served principle means that the last-arrived job is processed first, which is suitable for urgent order processing; The minimum slack time is to give priority to jobs close to the deadline, which is suitable for production environments that are sensitive to job deadlines and can accurately predict the remaining working time and current time; The critical ratio principle ensures that critical tasks are given priority by comprehensively considering task allowances and remaining working time. It is suitable for production environments that have strict requirements on task priority and can accurately predict task allowances and remaining working time.

6. The capacity balancing intelligent factory scheduling system based on MES according to claim 5 is characterized in that: The change scheduling module also includes: emergency shutdown rules for serious defects in quality inspection and emergency insertion rules for high-priority orders. The emergency shutdown rules for serious defects in quality inspection are that when the quality inspection link finds that the products produced by the production line equipment have serious defects, the operation of the relevant production line equipment is immediately stopped, and the production line equipment is troubleshooting and repair attempts are made; if the repair is ineffective, it is upgraded to a whole line shutdown, and the troubleshooting, repair and prevention mechanism is started; the emergency insertion rule for high-priority orders is that for orders set to the highest priority, the current production plan is adjusted, and the high-priority order is inserted at the top of the production line; the production of emergency orders is started, and the production progress and quality are continuously monitored.

7. The capacity balancing intelligent factory scheduling system based on MES according to claim 6 is characterized in that: The weight change of each production performance index according to the adopted change algorithm is specifically as follows: Dynamically calculate the weight of each indicator based on the entropy weight method-TOPSIS fusion model; Priority is sorted based on the weight of each indicator, and multi-objective optimization is performed through reinforcement learning algorithm to achieve the goal of minimizing entropy increase by changing the weight of each performance indicator.

8. The capacity balancing intelligent factory scheduling system based on MES according to claim 7 is characterized in that: The entropy weight method-TOPSIS fusion model dynamically calculates the weight of each indicator, specifically: Based on the dynamic correlation matrix, obtain the positive index and negative index of the dynamic correlation matrix; Based on the obtained indicators, calculate the proportion of each sample under each indicator to obtain the proportion matrix; Based on the weight matrix, the entropy value and redundancy of each indicator are obtained, and the weight of each indicator is finally determined; Based on the standardized dynamic correlation matrix and weight vector, a weighted decision matrix is ​​obtained; Based on the weighted decision matrix, positive ideal solutions and negative ideal solutions are determined; The distance between each indicator and the positive ideal solution and the negative ideal solution is calculated to obtain the comprehensive score of each indicator. The indicators are sorted according to the comprehensive score to obtain the final sorting result.

9. The capacity balancing intelligent factory scheduling system based on MES according to claim 8 is characterized in that: The positive index of the dynamic correlation matrix is ， The negative index of the dynamic correlation matrix is: ， in, is the element in the i-th row and j-th column of the original data matrix, is the minimum value in the jth column, is the maximum value in the jth column; Calculate the proportion of each sample under each indicator, specifically: ， in, is the sum of the standardized values ​​of all samples under this indicator; m is the number of rows and columns of the dynamic correlation matrix; The weighted decision matrix is ​​obtained based on the standardized dynamic association matrix and weight vector, which is specifically: Given the normalized dynamic incidence matrix and the weight vector w =( w 1, w 2,…, wn ), weighted decision matrix V The calculation formula is: ， in, The weight of each indicator; The distance between each indicator and the positive ideal solution and the negative ideal solution is calculated to obtain the comprehensive score of each indicator, which is specifically: The distance between each indicator and the positive ideal solution for ， The distance between each indicator and the negative ideal solution for ， in, is a positive ideal solution; is a negative ideal solution; The calculation formula of the comprehensive score is: 。 10. The capacity balancing intelligent factory scheduling system based on MES according to claim 9 is characterized in that: The performance indicators include: equipment utilization rate , Energy efficiency and changeover costs , and their corresponding weight coefficients α, β and γ; The weight of each performance indicator is changed by performing priority sorting based on the weight of each indicator and performing multi-objective optimization through a reinforcement learning algorithm, specifically: Set the state, action, reward and Q value update of the Q-learning algorithm; when Q-learning learns the optimal strategy, dynamically adjust the values ​​of α, β, and γ, and gradually achieve the required minimum entropy increase through iterative optimization; The state is the performance indicator value at the current moment ( , , ); the action is to adjust the values ​​of α, β, and γ; the Q value is updated by iteratively updating the Q value using the Q-learning update formula to find the optimal strategy, Among them, equipment utilization rate , Energy efficiency and changeover costs And the relationship between the corresponding weight coefficients α, β and γ is: ， in, is the equipment utilization rate, For energy efficiency, is the conversion cost, weight coefficient , and Through Q-learning dynamic adjustment, the established correlation matrix between the indicators is used to select and optimize data, and the required minimum entropy increase is achieved by adjusting the performance indicators.

Images