A real-time intelligent scheduling and management system for bamboo chopstick production line data

Abstract

The present invention relates to the technical field of production line scheduling management, and particularly to a real-time intelligent scheduling management system for bamboo chopstick production lines, including a data acquisition module, an intelligent scheduling module, an anomaly detection module, a device coordination and optimization module, and a decision support module; the data acquisition module collects data such as equipment operation status, production progress, raw material consumption, and production volume statistics; the intelligent scheduling module uses the federated learning algorithm to generate a production scheduling plan by integrating multiple factors; the anomaly detection module monitors anomalies with the help of a deep autoencoder; the device coordination and optimization module optimizes device coordination based on the former two; the decision support module generates a visual interface and an intelligent analysis report. The present invention realizes the real-time intelligent management of the production process, effectively improves production efficiency, optimizes resource utilization, guarantees product quality, facilitates management decision-making, promotes the intelligent development of the bamboo chopstick production industry, and enhances the competitiveness and sustainable development ability of enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of production line scheduling management, and specifically to a real-time intelligent scheduling management system for bamboo chopstick production line data. Background Art

[0002] In the field of bamboo chopstick production, with the continuous growth of market demand and increasing competition, improving production efficiency, reducing costs, and ensuring product quality have become important challenges faced by enterprises. The traditional management methods of bamboo chopstick production lines have many limitations and are difficult to meet the requirements of modern production;

[0003] Traditional scheduling methods often rely on manual experience and lack real-time and comprehensive mastery of data in multiple aspects such as equipment operation status, production progress, and raw material consumption. It is difficult for schedulers to make scientific and reasonable decisions quickly in a complex production environment, resulting in unbalanced production task allocation and low equipment utilization rate. For example, some equipment may be overloaded while other equipment remains idle for a long time, causing loss of production efficiency. At the same time, when facing sudden situations such as order changes and equipment failures, traditional scheduling methods cannot adjust the production plan in a timely and effective manner, easily leading to production chaos and affecting order delivery; managers lack sufficient data support and scientific analysis tools when making decisions. Traditional production management methods cannot provide comprehensive and accurate production data in a timely manner, and it is difficult for managers to deeply understand various problems and potential risks in the production process and make forward-looking decisions. For example, when deciding whether to increase production line equipment, adjust production processes, or purchase raw materials, due to the lack of detailed analysis and trend prediction of key indicators such as production efficiency, equipment utilization rate, and raw material consumption, decisions are often blind, which may lead to waste of resources or exacerbation of production bottlenecks. Therefore, in view of the above problems, a real-time intelligent scheduling management system for bamboo chopstick production line data is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a real-time intelligent scheduling management system for bamboo chopstick production line data to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A real-time intelligent scheduling management system for bamboo chopstick production line data, comprising:

[0007] A data acquisition module: It is arranged on the bamboo chopstick production line and is connected to sensors, PLC controllers, and production equipment interfaces for real-time acquisition of equipment operation status data, production progress data, raw material consumption data, and production volume statistical data during the production process of bamboo chopsticks;

[0008] Intelligent Scheduling Module: Communicates and connects with the data acquisition module. Based on the collected data of equipment operation status, production progress, raw material consumption, and production output statistics, it uses the federated learning algorithm. Considering the real-time status of equipment, current production requirements, and task priorities, it automatically generates a production scheduling plan for the bamboo chopstick production line through intelligent operations;

[0009] Abnormal Detection Module: Communicates and connects with the data acquisition module. With the help of the deep autoencoder AutoEncoder technology, it conducts real-time analysis on the collected data to monitor abnormal situations in the bamboo chopstick production process, including equipment failures, insufficient raw materials, and bottlenecks in production links;

[0010] Equipment Collaboration and Optimization Module: Communicates and connects with the intelligent scheduling module and the abnormal detection module. It uses the federated learning algorithm. Based on the production scheduling plan generated by the intelligent scheduling module and the abnormal situations feedback by the abnormal detection module, it optimizes the collaborative working mode among various equipment in the bamboo chopstick production line;

[0011] Decision Support Module: Communicates and connects with the intelligent scheduling module, the abnormal detection module, and the equipment collaboration and optimization module. Based on the data obtained from each module, it generates a visual interface for managers to view through big data analysis technology and regularly generates an intelligent analysis report covering production efficiency, equipment utilization rate, and raw material consumption.

[0012] As an optimal solution, the intelligent scheduling module includes:

[0013] Multi-objective Collaborative Optimization Unit: In the application of the federated learning algorithm, multiple equipment nodes or links participating in the bamboo chopstick production line scheduling are set as clients of the federated learning. The production efficiency is set as 、the equipment utilization rate is set as 、the energy consumption is set as 、the raw material waste is set as as the joint optimization objectives; each client constructs a local objective function based on the locally collected data , where, is the decision variable vector of client . The overall multi-objective optimization objective function is expressed as: , where, is the number of clients, is the weight of client . Through the federated learning mechanism, the model parameters are interacted and updated among clients to collaboratively optimize each objective;

[0014] The local objective function of client The construction method is as follows: with production efficiency For example, suppose the output of product per unit time is , the equipment running time is ,but , Represents the client production efficiency, Comprehensive consideration It is constructed with other local related factors and constraints according to preset rules and mathematical relationships to accurately reflect the local contribution to the overall multi-objective optimization;

[0015] Adaptive adjustment unit: collects production-related data from each client in real time and aggregates them into a real-time status data set for the production line , and set a predefined set of emergency adjustment rules ; When abnormal situations occur, such as urgent order insertion and sudden equipment failure, the global coordination mechanism based on federated learning is used. and , by updating the scheduling policy parameters of each client To adjust the production scheduling plan, the update process is expressed as ,in, is the adaptive adjustment function based on federated learning, is the updated scheduling strategy parameter;

[0016] For different types of abnormal situations, The corresponding specific adjustment strategy details are set in the system. In case of equipment failure, the system determines how to reallocate production tasks to other normal equipment according to the link where the faulty equipment is located and the remaining available equipment capacity. Through the coordination of federated learning, the adjustment strategy is sent to the corresponding client for execution.

[0017] Task priority determination unit: Based on the urgency of the delivery date of bamboo chopstick orders , Product type importance level and the adequacy of raw material supply To determine the priority of production tasks; each client Calculate the local priority score based on local relevant information , ,in, For Clients correspond , , The weight coefficient of

[0018] Global production task priority Through the aggregation mechanism of federated learning, the Obtained by using the weighted average method , where is the weight of the client in the global priority calculation.

[0019] As a preferred solution, the anomaly detection module includes:[[]]

[0020] Anomaly recognition and localization unit: Using a deep autoencoder to process the collected data, mapping the input data through the encoder Encoder into a hidden layer representation , and then reconstructing the output through the decoder Decoder , calculating the reconstruction error , setting a reasonable threshold , when , it is determined that there is an abnormal situation in the production link corresponding to the data;

[0021] Further analyze the feature distribution of the hidden layer representation , and combine the feature ranges of each device and each production link under normal conditions preset to locate the specific device or production link with anomalies;

[0022] Anomaly warning and feedback unit:

[0023] Once an abnormal situation is detected, immediately generate a warning signal including the anomaly type, occurrence location, and severity, and send it to the intelligent scheduling module and the device collaboration and optimization module through the communication link;

[0024] According to the severity of the anomaly, set different levels of warning prompt methods. For minor anomalies, only display prompt information on the visualization interface. For anomalies that seriously affect production, trigger both audible and visual alarms, and push the detailed anomaly information to the terminal devices of the management personnel in real time.

[0025] As a preferred solution, the device collaboration and optimization module includes:[[]]

[0026] Device - to - device task allocation optimization unit: Under the federated learning architecture, regard each device as a participating node. Based on the production scheduling plan generated by the intelligent scheduling module, combined with the current real - time status of each device and the anomaly information feedback by the anomaly detection module, construct an optimization model for device - to - device task allocation; Let the device set be , be the total number of devices, the production task set be , be the number of production tasks, define the allocation matrix , where the element represents whether the task is assigned to the device , the goal is to find the optimal allocation matrix , the optimization process is achieved by iteratively updating the allocation strategy among device nodes through the federated learning algorithm, that is, each device node adjusts its corresponding value, is the index of the device, is the index of the production task;

[0027] Considering the heterogeneity of devices, for each device set the task processing capacity vector , which represents its processing efficiency or capacity limit for different types of tasks. When optimizing task allocation, add a constraint condition to ensure that the total amount of tasks assigned to device does not exceed its processing capacity, that is , where, represents the workload of task , represents device 's processing capacity limit for the task belonging to the task type to ensure the rationality and feasibility of task allocation;

[0028] Device operating parameter collaborative adjustment unit: Each device shares its own operating parameters through the federated learning mechanism to build a collaborative adjustment model of operating parameters based on multiple devices; Let the operating parameter vector of device be , by analyzing the correlation between the operating parameters of different devices and their impact on the overall production;

[0029] Use the gradient descent optimization algorithm to update the operating parameters of each device in the federated learning framework. Let the learning rate be , in each iteration, the operating parameter update rule of device is , where, represents the updated operating parameter vector of device , is the learning rate, represents the overall production line performance loss function with respect to the operating parameter vector of device gradient, is the overall production line performance loss function

[0030] As a preferred solution, the decision support module includes:

[0031] Real-time Monitoring Visualization Unit: Construct a visualization interface to display the detailed status information of the bamboo chopstick production line in real time on the interface with intuitive charts, graphs, and data lists, covering equipment operation status, production progress, currently effective production scheduling plans, and equipment health status warning information;

[0032] Intelligent Report Generation Unit: According to the preset time period, use big data analysis technology to comprehensively analyze the historical production data and real-time data of the bamboo chopstick production line, and generate an intelligent analysis report covering production efficiency, equipment utilization rate, and raw material consumption;

[0033] In the intelligent analysis report, in addition to presenting the actual data and change trends of each indicator, data mining algorithms are also used to mine potential problems, optimization directions, and improvement suggestions existing in the production process.

[0034] As can be seen from the technical solutions provided by the present invention above, a real-time intelligent scheduling management system for a bamboo chopstick production line provided by the present invention has the following beneficial effects:

[0035] The intelligent scheduling module integrates multi-source data, uses the federated learning algorithm to achieve accurate task allocation and parameter setting, continuously optimizes the scheduling plan to adapt to production changes, reduces waiting time, and the equipment collaboration and optimization module, based on the scheduling plan and abnormal situations, optimizes the equipment collaboration mode through federated learning, improves equipment utilization rate, and realizes efficient production; the anomaly detection module uses a deep autoencoder to monitor data in real time, discovers and warns anomalies in a timely manner, quickly locates and solves problems, and prevents quality problems; intelligent scheduling and equipment collaboration optimization ensure the stable operation of equipment and the stability of process parameters, ensuring the consistency of product quality; the visualization interface of the decision support module displays the production status in real time, assisting management personnel to make decisions in a timely manner, and the interactive function facilitates in-depth data analysis; the regularly generated intelligent analysis report provides key indicators and optimization suggestions, helping management optimization, and the data can be used for strategic planning and performance evaluation; the system comprehensively optimizes scheduling considering multiple factors, adapts to complex working conditions and diverse production requirements, and the federated learning algorithm realizes the combination of global and local optimization, adapting to complex production environments. Brief Description of the Drawings

[0036] Figure 1 It is a schematic diagram of the overall structure of a real-time intelligent scheduling management system for a bamboo chopstick production line of the present invention. Detailed Embodiment

[0037] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0038] To better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific embodiments.

[0039] As Figure 1 shown, an embodiment of the present invention provides a real-time intelligent scheduling management system for bamboo chopstick production line data, including a data acquisition module, an intelligent scheduling module, an anomaly detection module, a device collaboration and optimization module, and a decision support module.

[0040] In this embodiment, the data acquisition module is set on the bamboo chopstick production line and is connected to sensors, PLC controllers, and production equipment interfaces to collect real-time data on the operating status of equipment, production progress, raw material consumption, and production volume statistics during the production process of bamboo chopsticks.

[0041] Furthermore, the specific operation steps of the data acquisition module are as follows:

[0042] 1. Module connection and deployment

[0043] Installation and configuration: Install the data acquisition module at a suitable position on the bamboo chopstick production line to ensure its compatibility with the operating environment of the production line, and perform necessary initialization configurations, such as setting communication parameters, data storage paths, etc.

[0044] Connect sensors:

[0045] Identify key sensors: Determine various sensors on the production line for monitoring equipment operating status, production progress, raw material consumption, and production volume statistics, such as temperature sensors, pressure sensors, position sensors, counters, etc.

[0046] Physical connection: Use appropriate cables to connect the output interfaces of the data acquisition module to various sensors to ensure a firm connection and stable signal transmission; for different types of sensors, corresponding adapters or interface modules may be required to achieve correct connection.

[0047] Connect PLC controller:

[0048] Determine the communication protocol: Understand the communication protocol used by the PLC controller on the production line. Common ones include Modbus, Profibus, Ethernet / IP, etc.

[0049] Establish a communication link: According to the selected communication protocol, use corresponding communication cables (such as Ethernet cables, serial cables, etc.) to connect the communication ports of the data acquisition module to the PLC controller; configure the same communication parameters as the PLC controller in the data acquisition module, such as IP address, port number, baud rate, data bits, stop bits, parity bits, etc., to ensure normal communication between the two.

[0050] Connect to the production equipment interface:

[0051] Analyze the equipment interface type: Check the interface forms of each device on the bamboo chopstick production line, which may include RS-232 / RS-485 serial ports, Ethernet interfaces, USB interfaces, or other special interfaces;

[0052] Adapter connection: For different interface types, select the appropriate connection method and conversion device (if necessary); for example, for a serial port interface, use a serial cable to connect directly; for an Ethernet interface, connect to the same local area network through a network cable; for a USB interface, use a USB data cable to connect and ensure that the device driver is correctly installed;

[0053] 2. Data acquisition and transmission

[0054] Data acquisition of equipment operating status:

[0055] Sensor data reading: The data acquisition module reads various physical quantity data during the operation of the equipment in real time according to the preset sampling period through the connection with the sensors, such as equipment temperature, vibration amplitude, motor speed, hydraulic pressure, etc.; these data can reflect the real-time working status of the equipment and provide a basis for equipment fault prediction and maintenance;

[0056] PLC data acquisition (optional): If some equipment operating status information is stored in the PLC controller, the data acquisition module reads the relevant status registers or variable values, such as the start / stop status of the equipment, fault alarm signals, operating modes, etc., through the communication with the PLC according to the agreed data format and address;

[0057] Data preprocessing: Perform preliminary processing on the acquired equipment operating status data, including data cleaning (removing outliers, noise, etc.), data conversion (converting analog quantities to digital quantities, performing physical quantity conversion according to sensor characteristics, etc.), and data formatting to make it meet the requirements of subsequent analysis and processing;

[0058] Data transmission: Transmit the processed equipment operating status data in real time to the intelligent scheduling module, anomaly detection module, and other relevant system modules through wired or wireless communication methods (such as Ethernet, Wi-Fi, 4G / 5G, etc.) for further analysis;

[0059] Data acquisition of production progress:

[0060] Identify progress monitoring points: Determine the key links and monitoring points in the bamboo chopstick production process that can accurately reflect the production progress, such as the input quantity of raw materials, the processing quantity of semi-finished products, the packaging quantity of finished products, etc.; install corresponding sensors at these monitoring points or use the counting function of the production equipment itself to obtain progress data;

[0061] Data acquisition and calculation: The data acquisition module obtains relevant data of the monitoring points in real time. For example, it counts the number of bamboo chopstick blanks passing through by means of an optoelectronic sensor installed on the conveyor belt, or reads the number of completed processing procedures from the control system of the production equipment, etc. According to the production process and flow, these data are comprehensively calculated to obtain accurate production progress information, such as the percentage of the number of completed products in the total order quantity, the progress percentage of the current production process, etc.

[0062] Progress data transmission: Transmit the calculated production progress data to the relevant modules in a timely manner so that the management can understand the operation status of the production line in real time and adjust the production plan and scheduling strategy in a timely manner.

[0063] Raw material consumption data acquisition:

[0064] Install raw material monitoring equipment (optional): If the production line does not have a built-in raw material consumption monitoring function, weighing sensors, liquid level sensors or flow sensors can be installed on the raw material storage containers (such as bins, storage tanks, etc.) to monitor the stock or usage of raw materials in real time.

[0065] Data acquisition and calculation: The data acquisition module regularly reads the data of the raw material monitoring equipment and calculates the raw material consumption rate and the cumulative consumption total per unit time by comparing with the initial raw material quantity. For some production links with batch input of raw materials, the raw material consumption can also be calculated by recording the raw material input batches and the quantity of each batch.

[0066] Data transmission and storage: Transmit the raw material consumption data to the corresponding system modules for analysis and processing, and store it in the local database or cloud server for subsequent query and statistical analysis to provide data support for the raw material procurement plan.

[0067] Output statistics data acquisition:

[0068] Finished product counting and statistics: Install a counter at the finished bamboo chopstick packaging link or the final output position, or use the counting function of the automated packaging equipment to count the finished bamboo chopsticks in real time. The data acquisition module obtains the counter data and records the output of finished products in each time period.

[0069] Data integration and transmission: Integrate the output statistics data with the production progress data, equipment operation status data, etc. to form a complete production data record. Transmit the output statistics data to the decision support module through network communication for generating production reports, analyzing production efficiency and evaluating the completion of production tasks.

[0070] 3. Data verification and storage (optional)

[0071] Data verification:

[0072] Integrity check: Check whether the collected data is complete, ensure that each data item has a corresponding collected value without omission; for missing data, according to the importance and recoverability of the data, take corresponding processing measures, such as supplementing default values, using interpolation algorithms for estimation, or marking as abnormal data for subsequent processing;

[0073] Accuracy check: Verify the accuracy of the collected data by comparing it with preset reasonable data ranges, thresholds, or historical data; for data beyond the normal range, conduct secondary confirmation or judge its validity through cross-verification of multiple data sources; if data anomalies are found, issue an alarm in a timely manner and notify relevant personnel for inspection and processing;

[0074] Consistency check: Check whether there are logical contradictions or inconsistencies between data from different sources or at different collection times; for example, the operating status data collected by different sensors of the same device should match each other, and the production progress data should be logically consistent with the raw material consumption data and production statistics data; if inconsistencies are found, further analyze the reasons and make data corrections or adjustments;

[0075] Data storage:

[0076] Local storage: Store the verified collected data in the local storage device (such as hard disk, SD card, etc.) of the data acquisition module, organize it according to a certain data structure and file format for subsequent quick query and retrieval; at the same time, set data storage period and capacity management policies. When the storage device space is insufficient, delete old data or perform data backup according to certain rules;

[0077] Remote storage (optional): To ensure data security and scalability, the collected data can be transmitted over the network to a remote server or cloud storage platform for backup and storage; this can not only prevent local data loss, but also facilitate data sharing and centralized management among multiple production lines or factories, providing support for enterprise-level data analysis and decision-making; during data transmission, encryption technology is used to ensure data security and prevent data leakage;

[0078] 4. System monitoring and maintenance

[0079] Operating status monitoring: Establish an operating status monitoring mechanism for the data acquisition module to monitor the working status of the module in real time, including the operating conditions of hardware devices (such as whether sensors are working properly, whether communication links are unobstructed, whether storage devices are available, etc.) and the running status of software programs (such as whether data acquisition processes are running normally, whether data transmission is timely and accurate, etc.); through the monitoring interface or log files, any abnormal situations or fault alarm information can be found and displayed in a timely manner;

[0080] Fault diagnosis and repair: When the monitoring system detects an anomaly or a fault, it automatically starts the fault diagnosis program to analyze and troubleshoot possible causes of the fault; based on the diagnosis results, it provides corresponding solutions and operation suggestions, such as restarting the device, replacing faulty components, adjusting communication parameters, etc.; for some common faults, the system can automatically perform repair operations; for more complex faults, it promptly notifies the maintenance personnel to handle them on-site to ensure that the data acquisition module can quickly resume normal operation and reduce the data acquisition interruption time caused by faults;

[0081] Regular maintenance and calibration: Develop a regular maintenance plan for the data acquisition module, including cleaning, maintenance, inspection, and calibration of hardware devices, as well as software program upgrades and optimizations; regularly calibrate the sensors to ensure that their measurement accuracy meets production requirements; check the connection status of communication cables and interfaces to prevent loosening or damage from affecting data transmission; update the software of the data acquisition module to fix possible vulnerabilities and improve system performance and stability; during the maintenance process, record the maintenance operations and changes in device status in detail to provide a reference basis for subsequent maintenance decisions and fault analysis.

[0082] In this embodiment, the intelligent scheduling module is communicatively connected to the data acquisition module. Based on the collected device operation status data, production progress data, raw material consumption data, and production output statistics data, using the federated learning algorithm, and comprehensively considering the real-time status of the devices, current production requirements, and task priorities, it automatically generates a production scheduling plan for the bamboo chopstick production line through intelligent calculations. The intelligent scheduling module includes:

[0083] Multi-objective collaborative optimization unit: In the application of the federated learning algorithm, set multiple device nodes or links participating in the scheduling of the bamboo chopstick production line as clients of the federated learning, and set production efficiency as 、equipment utilization rate as 、energy consumption as 、raw material waste as as the joint optimization objectives; each client based on the locally collected data constructs a local objective function , where is the decision variable vector of client . The overall multi-objective optimization objective function is expressed as: , where is the number of clients, is the weight of client . The model parameters are interacted and updated among the clients through the federated learning mechanism to collaboratively optimize each objective;

[0084] The local objective function of client ​ The construction method is as follows: taking production efficiency as an example, assume that the product output per unit time is , and the equipment operation time is , then , represents the production efficiency of the client . Considering and other local relevant factors and constraints comprehensively, it is constructed according to the preset rules and mathematical relationships to accurately reflect the local contribution to the overall multi-objective optimization;

[0085] Adaptive adjustment unit: Collect production-related data feedback from each client in real time, and summarize it to form a real-time status data set of the production line , and at the same time set a predefined emergency adjustment rule set ; when abnormal situations such as emergency order insertion and sudden equipment failure occur, based on the global coordination mechanism of federated learning, according to and , adjust the production scheduling plan by updating the scheduling policy parameters of each client. The update process is expressed as , where is the adaptive adjustment function based on federated learning, and is the updated scheduling policy parameter

[0086] For different types of abnormal situations, corresponding specific adjustment policy details are set in . In case of equipment failure, determine how to re-allocate production tasks to other normal equipment according to the link where the faulty equipment is located and the remaining available equipment capacity. Through the coordination of federated learning, the adjustment policy is sent to the corresponding client for execution;

[0087] Task priority determination unit: Determine the production task priority according to the urgency of the delivery date of the bamboo chopstick order , the importance level of the product type and the adequacy of raw material supply ; each client calculates the local priority score according to the local relevant information , , where is the weight coefficient corresponding to the client for , , ;

[0088] The global production task priority Through the aggregation mechanism of federated learning, the is obtained by using the weighted average method , where is the weight of client in the global priority calculation;

[0089] Furthermore, the specific operation steps of the intelligent scheduling module are as follows:

[0090] 1. Data collection and communication

[0091] Establish a stable connection: The intelligent scheduling module uses Ethernet, Wi-Fi or other applicable communication technologies to build a reliable communication link with the data collection module to ensure the stability and real-time nature of data transmission;

[0092] Real-time data acquisition: Continuously receive device operation status data (such as device temperature, rotation speed, operation duration, etc.), production progress data (the number of completed items in each process, remaining workload, etc.), raw material consumption data (real-time raw material inventory, consumption rate, etc.) and production volume statistics data (the number of products already produced, production volume of different specifications, etc.) from the data collection module;

[0093] Data caching and sorting: Temporarily store the received data in the local buffer area and classify and sort it according to data types, time series, etc., for convenient subsequent processing and analysis;

[0094] 2. Data preprocessing and feature engineering

[0095] Data cleaning:

[0096] Remove possible error data points during the collection process, such as values that are significantly outside the normal range (for example, the device temperature is higher than the limit value that the device can withstand);

[0097] Handle missing data. According to the importance and distribution of the data, select a suitable method for filling, such as using the mean, median or specific interpolation algorithms;

[0098] Data transformation and normalization:

[0099] Unify the conversion of data with different units and magnitudes. For example, convert the device operation time to a unified time format and convert the raw material consumption data to a unified mass or volume unit;

[0100] Normalize the data so that each feature value is at the same order of magnitude. Commonly used methods include min-max normalization or standardization to improve the calculation efficiency and accuracy of subsequent algorithms;

[0101] Feature extraction and construction:

[0102] Mine key features related to production scheduling from the original data, such as the average load rate of equipment (calculated by the operating power and rated power of the equipment), the production rhythm of each process on the production line (the time interval between the completion of two adjacent products for this process), the raw material inventory turnover rate (the ratio of the raw material consumption speed to the average inventory level), the month-on-month growth rate of production volume, etc.;

[0103] According to the production process and business logic, construct new features. For example, combine the equipment operating status and production progress data to calculate the utilization trend of the equipment in different production stages; according to the raw material consumption and production volume data, estimate the unit raw material consumption of the product and its change trend to provide richer information for subsequent analysis and decision-making;

[0104] 3. Construction and Initialization of the Federated Learning Model

[0105] Model architecture selection: According to the characteristics and optimization objectives of the bamboo chopstick production line, select a suitable federated learning model architecture, such as a distributed neural network, a federated decision tree, etc.; taking the distributed neural network as an example, determine hyperparameters such as the number of network layers, the number of neurons in each layer, and the activation function, and construct the basic model structure;

[0106] Client model initialization: Initialize the selected model architecture on each device node (client) participating in the federated learning, and assign the same initial model parameters to each client to ensure that the model has a consistent starting point globally, which is convenient for subsequent collaborative training and optimization;

[0107] 4. Multi-objective Collaborative Optimization

[0108] Objective function setting:

[0109] Clarify the joint optimization objectives, such as production efficiency (P), equipment utilization rate (U), energy consumption (E), raw material waste (W);

[0110] For each objective, combine the preprocessed data and the actual production situation to construct a specific objective function expression; taking production efficiency as an example, production efficiency can be expressed as the ratio of the output quantity of qualified products within a unit time to the equipment operating time, that is , where is the quantity of qualified products, is the equipment operating time; at the same time, consider the impact of factors such as equipment failures, maintenance plans, and process adjustments on production efficiency, and incorporate these factors into the constraint conditions of the objective function;

[0111] Assign weights to the objective functions of each client ( ), the determination of weights is based on the importance that the enterprise attaches to each goal and the importance of the client in the production process; for example, for the equipment client in the key production link, the weights of the production efficiency and equipment utilization rate goals may be relatively high; while for the equipment client with large energy consumption, the weight of the energy consumption goal is more prominent;

[0112] Local model training:

[0113] Each client, based on the locally collected data ( ), uses the local objective function ( ) to train the local model; during the training process, the client calculates the error (loss function) between the model prediction value and the actual value according to its own data characteristics and the objective function, and adjusts the parameters of the local model ( ) through the backpropagation algorithm or other optimization algorithms to minimize the loss function;

[0114] During the training process, the client can adopt optimization algorithms such as Stochastic Gradient Descent (SGD), Adagrad, Adadelta, etc., select appropriate learning rates and iteration times according to the data distribution and the complexity of the model, ensure that the local model can effectively learn the patterns and rules in the local data, and avoid the occurrence of overfitting;

[0115] Model parameter uploading and aggregation:

[0116] After a certain number of rounds of local training, each client uploads the parameters obtained from the local model training (such as the weight matrix and bias vector in the neural network) to the central server or the coordinator node;

[0117] The central server uses the Federated Averaging (FedAvg) algorithm or other appropriate aggregation algorithms to aggregate the uploaded client parameters; during the aggregation process, the parameters of each client are weighted and averaged according to the weights of the clients ( ) to obtain the updated parameters of the global model; for example, for the weight parameters of the neural network , the globally updated weight The calculation formula is: , where is the number of clients, is the weight parameter uploaded by the client;

[0118] Global model parameter distribution and update:

[0119] The central server distributes the aggregated global model parameters back to each client;

[0120] The client updates the local model using the received global model parameters to synchronize the local model with the global model, achieving collaborative optimization of the model on a global scale;

[0121] By iterating the above processes of local training, parameter uploading, aggregation, and download update multiple times, the model continuously learns and adapts to various data patterns on the production line, gradually optimizing each objective function until the preset convergence condition is reached (such as the loss function value is lower than a certain threshold or the number of iterations reaches the upper limit). At this time, the obtained model parameters are the results of multi-objective collaborative optimization for the bamboo chopstick production line;

[0122] 5. Adaptive adjustment strategy

[0123] Real-time data monitoring and collection:

[0124] The intelligent scheduling module continuously monitors the real-time status of the production line and collects the latest data feedback from each client, including sudden equipment failure information (failed equipment number, failure type, occurrence time, etc.), emergency order insertion situations (order number, product specifications, quantity, delivery date, etc.), real-time changes in equipment operation parameters (such as sudden increase or decrease in equipment load), and dynamic changes in raw material supply (such as raw material inventory being lower than the safety threshold, delay in arrival of new raw materials, etc.);

[0125] Integrate this real-time data into the production line real-time status data set ( ) to provide timely and accurate data support for subsequent adaptive adjustments;

[0126] Emergency adjustment rule matching:

[0127] According to the preset emergency adjustment rule set ( ), quickly match and identify abnormal situations in the real-time status data set ( ); The rule set ( ) formulates detailed coping strategies and adjustment rules for different types of abnormal situations. For example:

[0128] When a device fails, the rule stipulates that according to the process where the failed device is located, the processing capacity of the remaining available devices, and the task priority, reassign the tasks on the failed device to other normal devices; at the same time, adjust the operation parameters of the relevant devices to meet the requirements of the new tasks and ensure the continuity of the production process;

[0129] For the insertion of emergency orders, the rule determines that according to the priority, product type, and delivery date of the order, adjust the production plan without affecting the production progress of the original orders, and give priority to arranging the production of emergency orders; this may involve measures such as reallocating raw materials, adjusting equipment schedules, and optimizing production processes to ensure that emergency orders can be delivered on time;

[0130] Calculation of scheduling policy parameter update:

[0131] When an abnormal situation is recognized and the corresponding adjustment rule is matched, based on the global coordination mechanism of federated learning, the new scheduling policy parameters of each client are calculated through a specific adaptive adjustment function (g) ( ); The adjustment function (g) determines how to adjust the policy parameters such as production task allocation and equipment operation parameters for each client under abnormal situations by considering various factors, such as the remaining available time of the equipment, the urgency of the task, and the raw material inventory level, according to the real-time status data set ( ), the emergency adjustment rule set ( ), and the current client's scheduling policy parameters ( );

[0132] For example, in the case of equipment failure, the adjustment function (g) may calculate parameters such as the proportion of tasks associated with the faulty equipment that should be assigned to other equipment and the adjustment coefficient of the operating speed of other equipment to achieve a reasonable transfer of production tasks and a balanced operation of the production line;

[0133] Execution and feedback of adjustment strategy:

[0134] The updated scheduling policy parameters ( ) are sent to the corresponding client, and the client executes the adjusted production scheduling policy according to the new parameters;

[0135] During the process of the client executing the adjustment strategy, the execution situation and production data are fed back to the intelligent scheduling module in real time, including the actual operating status of the equipment, the task completion progress, the raw material consumption situation, etc.; The intelligent scheduling module continuously monitors the implementation effect of the adjustment strategy based on the feedback information and judges whether the production process has returned to normal or is developing towards the expected goal;

[0136] If it is found that there are still problems in the adjusted production process or the expected effect is not achieved, the intelligent scheduling module further optimizes the adjustment strategy according to the actual situation, updates the scheduling policy parameters again until the production line can operate stably or meet the emergency production requirements under abnormal situations; At the same time, record and analyze the entire adaptive adjustment process, summarize experience and lessons, provide a reference basis for handling similar abnormal situations in the future, and continuously improve the emergency adjustment rule set ( ) and the adaptive adjustment mechanism;

[0137] 6. Determination of task priority:

[0138] Quantitative evaluation of priority factors

[0139] Urgency of order delivery date ( ): Calculate the remaining time for each order until the delivery date. The shorter the remaining time, the higher the urgency level of the delivery date. The reciprocal of the time difference or other appropriate functions can be used to convert the remaining time into a priority score. For example, , where is the remaining delivery time of the order. At the same time, considering the production cycle of the enterprise and the importance level of the order, differential weighting is performed on the urgency level of the delivery date for different types of orders to ensure that key orders receive higher priority attention;

[0140] Product type importance level ( ): Set the importance level values for different types of bamboo chopstick products according to the enterprise's product strategy and market demand. For example, the importance level of high-value-added and customized products can be set to a higher value, while the importance level of ordinary standard products is set to a lower value. In addition, considering factors such as the profit contribution of the product and the stability of market demand, the importance level is dynamically adjusted to accurately reflect the importance of the product at different times;

[0141] Raw material supply sufficiency situation ( ): Real-time monitor the raw material inventory level and supply stability, and calculate the available days of raw material inventory ( , where is the current raw material inventory quantity, and is the average daily consumption rate of raw materials); when the available days of raw material inventory are lower than the safety threshold, it indicates that the raw material supply is tight, and the priority of the corresponding production tasks should be increased; at the same time, considering the uncertainty factors of raw material supply (such as the risk of supplier delivery delay, raw material market price fluctuations, etc.), a comprehensive evaluation of the raw material supply sufficiency situation is carried out to ensure that when problems may occur in raw material supply, relevant production tasks are arranged first to avoid production interruption due to raw material shortage;

[0142] Local priority score calculation:

[0143] Each client calculates the local priority score ( ) according to the local production task information and the above quantitatively evaluated priority factors; the calculation formula is , where , , are the weight coefficients corresponding to the client for , , ; these weight coefficients are set according to the characteristics of the production links where the clients are located and the enterprise's production management strategy. For example, for clients that are large consumers of raw materials, may have a higher weight; for clients that produce high-value-added products, has a relatively larger weight;

[0144] When calculating the local priority score, fully consider the impact of factors such as the client's local device status, task progress, and personnel configuration on the priority, ensuring that the priority score can accurately reflect the actual urgency and importance of local production tasks;

[0145] Global priority aggregation and adjustment:

[0146] The intelligent scheduling module collects the local priority scores of each client through the aggregation mechanism of federated learning ( ), and calculates the global production task priority using the weighted average method ( ). The calculation formula is , where is the weight of the client in the global priority calculation; the weight is determined by comprehensively considering factors such as the importance of the client in the production line, data quality, and historical task completion, ensuring that the global priority can reasonably reflect the production requirements and task priority distribution of the entire production line;

[0147] Sort and adjust the production tasks according to the global priority, give priority to arranging tasks with high priority for production, and at the same time consider the relevance between tasks and resource constraints to ensure the efficient and orderly production process; during the production process, dynamically update the task priority according to the actual situation. For example, when an emergency order is inserted or there is a sudden change in raw material supply, re-evaluate and adjust the task priority in a timely manner to adapt to the changes in the production environment;

[0148] 7. Generation and issuance of production scheduling plan

[0149] Task allocation and sorting based on the optimization results:

[0150] Based on the results determined by the equipment operation parameters and task priorities obtained from the multi-objective collaborative optimization, allocate the most suitable equipment and execution time for each production task; during the task allocation process, fully consider the real-time status of the equipment (including whether the equipment is available, the equipment load situation, the equipment maintenance plan, etc.), the processing capacity of the equipment (such as the processing efficiency of the equipment for different types of tasks, the maximum processing capacity of the equipment, etc.), and the task priority, and use greedy algorithms, dynamic programming algorithms or other appropriate task allocation algorithms to reasonably allocate tasks to each equipment to achieve the goals of maximizing production efficiency, equalizing equipment utilization, and minimizing energy consumption and raw material waste;

[0151] At the same time, the production tasks are sorted and the execution order of the tasks is determined; the sorting process comprehensively considers the production process constraints (such as certain processes must be carried out in a specific order), the equipment switching cost (the time and resources required for the equipment to switch from one task to another) and the correlation between tasks (such as the material transfer efficiency and time interval between the previous and next processes); by reasonably arranging the task sequence, the waiting time, equipment idle time and material handling time in the production process are reduced, and the overall production efficiency and smoothness of the production line are improved;

[0152] Generate detailed scheduling instructions:

[0153] Convert the determined task allocation and sorting results into specific scheduling instructions, including equipment start / stop instructions, equipment operating parameter setting instructions (such as equipment operating speed, temperature, pressure and other process parameters), task start / end time notifications, material delivery instructions (including raw material delivery quantity, delivery time and delivery destination, etc.) and personnel deployment instructions (such as arranging operators to designated equipment for production operations), etc.;

[0154] The format and communication protocol of the dispatching instructions should be compatible with the equipment control system and operator terminals on the production line to ensure that the instructions can be accurately transmitted and executed;

[0155] Plan issuance and execution monitoring:

[0156] The generated dispatching instructions are sent to each equipment node, material distribution system and operator terminal in real time through the communication network to ensure that each link in the production line can receive the instructions in time and execute the production tasks as required;

[0157] The intelligent scheduling module establishes a production process monitoring mechanism, tracks the execution of the scheduling plan in real time, collects feedback information such as equipment operation data, production progress data, material consumption data, etc., compares and analyzes it with the preset production plan and scheduling plan, and promptly discovers deviations and abnormalities in the execution process;

[0158] When deviations are found between the production process and the scheduling plan, such as equipment failures causing task delays, insufficient material supply affecting production progress, etc., the corresponding adjustment mechanism is immediately activated to dynamically optimize and adjust the scheduling plan according to the real-time situation, and the adjusted scheduling instructions are re-issued to ensure that the production process always proceeds smoothly towards the predetermined goal; at the same time, the execution status and adjustment process of the production scheduling plan are recorded and analyzed to provide data support and experience reference for the subsequent optimization and improvement of production scheduling;

[0159] 8. Solution evaluation and continuous optimization

[0160] Data collection and indicator calculation:

[0161] After the production scheduling plan has been implemented for a period of time, collect the actual production data during this period, including equipment operation data (such as equipment utilization rate, equipment failure rate, equipment operation time, etc.), production progress data (actual completion time of each process, product output, production cycle, etc.), raw material consumption data (total actual raw material consumption, raw material utilization rate, raw material waste rate, etc.), and order delivery data (order on-time delivery rate, number of delayed orders and delay time, etc.);

[0162] According to the collected data, calculate the evaluation indicators related to production scheduling, such as production efficiency (the ratio of actual output to theoretical maximum output), overall equipment effectiveness (the ratio of the actual operation time of each equipment to the total available operation time), energy consumption rate (the ratio of actual energy consumption to production output), raw material waste rate (the ratio of raw material waste to total raw material input), order delivery on-time rate (the ratio of the number of orders delivered on time to the total number of orders), etc.; these indicators comprehensively reflect the implementation effect of the production scheduling plan and the operation performance of the production line;

[0163] Scheme effect evaluation and analysis:

[0164] Compare and analyze the calculated evaluation indicators with the preset target values or historical data to evaluate the advantages and disadvantages of the production scheduling plan; if all indicators meet or exceed the expected targets, it indicates that the current scheduling plan is relatively effective; if some indicators do not meet the expectations, such as low production efficiency, uneven equipment utilization, high energy consumption, or delayed order delivery, etc., it is necessary to deeply analyze the reasons and find out the key factors affecting the production scheduling effect;

[0165] Adopt data analysis methods (such as data mining, statistical analysis, etc.) to deeply mine the production data, analyze the correlation and influence degree between different factors, such as the relationship between equipment operation parameters and product quality, the relationship between production task allocation and equipment utilization, the relationship between raw material supply and production progress, etc.; through data analysis, determine the root causes of production scheduling problems and provide a basis for subsequent optimization and improvement;

[0166] Model and strategy optimization and adjustment:

[0167] According to the evaluation and analysis results, optimize and adjust the federated learning model and production scheduling strategy; if it is found that the model is inaccurate in prediction or unable to effectively optimize the objective function in some cases, it may be necessary to adjust the model structure (such as increasing or decreasing the number of neural network layers, adjusting the number of neurons, etc.), hyperparameters (such as learning rate, number of iterations, regularization coefficient, etc.), or optimization algorithms; at the same time, check the data preprocessing and feature engineering steps to ensure the data quality and feature effectiveness of the input model;

[0168] For production scheduling strategies, based on the causes of the problems analyzed, targeted improvements can be made to the task allocation algorithm, priority determination rules, equipment collaborative working mode, etc. For example, if it is found that some equipment often has a backlog of tasks, resulting in excessive utilization, while other equipment has a low utilization rate, it may be necessary to adjust the task allocation strategy to distribute tasks more evenly; if the order delivery timeliness rate is low, it may be necessary to optimize the task priority determination method to highlight the impact of the urgency of the delivery date on the priority, or adjust the production plan to reserve enough production time in advance to deal with urgent orders;

[0169] Optimization solution implementation and verification:

[0170] Apply the optimized federated learning model and production scheduling strategy to actual production and implement a new round of production scheduling. During the implementation process, pay close attention to changes in the production process, collect relevant data, observe the changing trends of various evaluation indicators, and verify the effectiveness of the optimization measures.

[0171] If the optimization plan fails to achieve the expected results, it is necessary to re-evaluate and analyze, further identify problems, and adjust the optimization strategy until the best solution is found; through continuous evaluation, optimization, implementation and verification, the scientificity and rationality of the production scheduling plan will be continuously improved, and the overall operational efficiency and competitiveness of the bamboo chopstick production line will be improved;

[0172] Knowledge accumulation and experience inheritance:

[0173] In the process of program evaluation and continuous optimization, we should focus on accumulating knowledge and experience in production data, optimization strategies, problem-solving methods, etc.; establish a knowledge base or case library to organize and archive successful optimization cases and effective solutions so that they can be quickly referenced and learned when encountering similar problems in the future;

[0174] Regularly organize training and exchanges for production management personnel, technical personnel and operators to share experiences and lessons learned in the optimization process, improve the team's overall production scheduling management level and problem-solving ability; promote continuous improvement and innovative development of the company's production management through knowledge accumulation and experience inheritance.

[0175] In this embodiment, the abnormality detection module is connected to the data acquisition module in communication, and the collected data is analyzed in real time by using the deep autoencoder technology to monitor abnormal conditions in the production process of bamboo chopsticks, including equipment failure, insufficient raw materials, and bottlenecks in the production process. The abnormality detection module includes:

[0176] Anomaly recognition and positioning unit: Use deep autoencoders to process the collected data and convert the input data Mapped to hidden layer representation through encoder , and then reconstructed by the decoder Decoder to obtain the output , calculate the reconstruction error , set a reasonable threshold , when , it is determined that there is an abnormal situation in the production link corresponding to the data;

[0177] Further analyze the feature distribution of the hidden layer representation , and combine the feature ranges in the normal state of each device and each production link preset to locate the specific device or production link where the abnormality occurs;

[0178] Abnormal warning and feedback unit:

[0179] Once an abnormal situation is detected, an early warning signal including the abnormal type, occurrence location, and severity is immediately generated and sent to the intelligent scheduling module and the device collaboration and optimization module through the communication link;

[0180] According to the severity of the abnormality, different levels of early warning prompt methods are set. For minor abnormalities, only prompt information is displayed on the visualization interface. For abnormalities that seriously affect production, audible and visual alarms are triggered simultaneously, and the detailed abnormal information is pushed to the terminal devices of the management personnel in real time;

[0181] Furthermore, the specific operation steps of the abnormal detection module are as follows:

[0182] 1. Data collection and transmission

[0183] Establish a communication link: The abnormal detection module establishes a stable communication connection with the data collection module through Ethernet, Wi-Fi or other reliable communication methods to ensure that data can be transmitted in real time and accurately;

[0184] Receive the collected data: Continuously obtain the equipment operation status data (such as equipment temperature, vibration, pressure, current, rotation speed, etc.), production progress data (quantity completed in each process, product transfer time, etc.), raw material consumption data (raw material inventory change, consumption rate, etc.), and production volume statistical data (quantity of finished products produced, production volume of different specifications of products, etc.) during the production process of bamboo chopsticks from the data collection module;

[0185] Data caching and preprocessing: The received data is temporarily stored in the local buffer area for preliminary data preprocessing, including data cleaning (removing noise and outliers), data format conversion (unifying data types and encoding formats), data normalization (mapping data to a specific interval, such as [0,1] or [-1,1], to improve the efficiency of model training and analysis), etc., so that the data meets the input requirements of the deep autoencoder model;

[0186] 2. Construction and training of the deep autoencoder model

[0187] Model Architecture Design:

[0188] Determine the network structure of the deep autoencoder, including the number and number of nodes in the input layer, hidden layer, and output layer; determine the number of input layer nodes according to the number of features of the collected data. For example, if 10 parameters of the device operating state are collected, the number of input layer nodes is 10; the design of the hidden layer is determined according to the complexity of the data and the expression ability of the model. Generally, a multi-layer structure can be adopted, such as 2-5 layers, and the number of nodes in each layer gradually decreases to form a "funnel" shape to extract the key features of the data; the number of output layer nodes is the same as that of the input layer, which is used to reconstruct the input data;

[0189] Select a suitable activation function. For example, the ReLU (Rectified Linear Unit) function is used for the hidden layer to introduce non-linearity and enhance the expression ability of the model; the output layer can select a suitable activation function according to the needs of reconstructing the data. For example, a linear activation function is used for regression tasks (such as reconstructing device operating state data), or the Sigmoid function is used to map the output value to the interval [0,1] (such as for judging the probability of anomalies);

[0190] Model Training Preparation:

[0191] Divide the preprocessed historical production data into a training set and a validation set. Usually, the training set is allocated according to a ratio of 70%-80%, and the remaining part is the validation set; ensure that the training set can fully represent the data pattern under normal production conditions and cover data samples under various different working conditions and production conditions;

[0192] Define a loss function to measure the difference between the model-reconstructed data and the original data; commonly used loss functions include the mean square error (MSE) function. For multivariate data, MSE can calculate the average of the reconstruction errors of each variable. The formula is , where is the number of samples, is the original data sample, is the data sample reconstructed by the model;

[0193] Model Training Process:

[0194] The deep autoencoder model is trained using the training set, and the weights and bias parameters of the model are adjusted through the backpropagation algorithm to minimize the loss function. During the training process, appropriate training parameters such as the learning rate (e.g., 0.001 - 0.01), the number of iterations (e.g., 1000 - 5000 times), and the batch size (e.g., 32 - 128) are set. The learning rate determines the step size of each parameter update. If it is too large, the model may not converge; if it is too small, the training process will be too long. The number of iterations controls the number of training rounds until the loss function value on the training set no longer decreases significantly or reaches the preset iteration upper limit. The batch size affects the stability and efficiency of model training. An appropriate batch size can accelerate the training process and improve the generalization ability of the model.

[0195] After each round of training, the model is evaluated using the validation set, and the loss function value and other evaluation metrics (such as accuracy, recall, etc., if applicable) on the validation set are calculated. Observe the change trend of the validation set metrics to determine whether the model has overfitting or underfitting phenomena. If the loss function value on the validation set continues to decrease, it means that the model is still learning effective features and has not reached the best state. If the loss function value on the validation set starts to rise while the loss function value on the training set continues to decrease, overfitting may occur. The model begins to overlearn the detailed features of the training set and loses the generalization ability for new data. At this time, measures such as early stopping of training, increasing regularization terms (such as L1 or L2 regularization), or adjusting the model structure can be taken to prevent overfitting.

[0196] After multiple trainings and validations, until the performance of the model on the validation set reaches stability and meets the expected requirements, the training process of the model is completed, and the trained deep autoencoder model is obtained.

[0197] 3. Real - time Data Monitoring and Anomaly Identification

[0198] Data Input and Reconstruction:

[0199] The pre - processed production data collected in real - time is input into the trained deep autoencoder model. Through the encoder part of the model, the input data is mapped to the hidden layer representation ( ), and then through the decoder part, the output data is reconstructed ( ).

[0200] Reconstruction Error Calculation and Anomaly Judgment:

[0201] Calculate the reconstruction error ( ) between the reconstructed data and the original input data, that is, calculate the square root of the sum of the squares of the differences between the two in each dimension. Set a reasonable threshold , the threshold can be determined by analyzing the distribution of reconstruction errors on the training set and the validation set. For example, the 95th percentile or 99th percentile of the reconstruction error distribution of normal data can be taken as the threshold; when the reconstruction error is less than, it is determined that there may be an abnormal situation in the production link corresponding to the data;

[0202] Abnormality location and analysis:

[0203] For the data determined to be abnormal, further analyze the feature distribution of the model's hidden layer representation ; Visualization techniques (such as drawing scatter plots after PCA dimensionality reduction of the main components, drawing feature heat maps, etc.) can be used to observe the change patterns of the hidden layer features. Combining with the feature ranges of each device and each production link under normal conditions preset, locate the specific device or production link where the abnormality occurs; for example, if it is found that the values of several features corresponding to a certain device in the hidden layer representation deviate significantly from the normal range, and the production link related to this device is in a key position in the production process, then it can be preliminarily judged that there is an abnormality in this device or its production link; at the same time, compare the differences between the abnormal data and the normal data in each feature of the input layer, and analyze which parameter changes have a greater impact on the reconstruction error, so as to more accurately determine the root cause of the abnormality. For example, too high device temperature may cause significant changes in the feature representation of the device operation state data in the model, thereby causing a large reconstruction error;

[0204] 4. Abnormality early warning and feedback

[0205] Early warning signal generation:

[0206] Once an abnormal situation is detected, the abnormality detection module immediately generates an early warning signal including the type of abnormality (such as equipment failure, insufficient raw materials, production link bottleneck, etc.), the location of occurrence (specific equipment number, production process name, etc.), and the severity level (classified according to the size of the reconstruction error or other relevant indicators, such as minor, moderate, severe); the early warning signal adopts a standardized data format for communication and interaction with other system modules;

[0207] Early warning information transmission:

[0208] The early warning signal is sent to the intelligent scheduling module and the equipment coordination and optimization module through the communication link, so that these modules can obtain the abnormal information in time and take corresponding countermeasures; at the same time, the early warning information is displayed on the display screen or monitoring terminal set at the production site to remind the on-site operators to pay attention to the occurrence of abnormal situations;

[0209] Early warning prompt method setting:

[0210] Set different levels of warning prompt methods according to the severity of the anomaly; for minor anomalies (such as slight fluctuations in equipment operating parameters but still within the acceptable range), only display prompt information on the visualization interface to attract the attention of relevant personnel, but do not affect the normal production process; for anomalies that seriously affect production (such as key equipment failures, severe shortages of raw materials, etc.), trigger audible and visual alarms simultaneously, and push the detailed anomaly information (including details of anomaly data, possible impact scope, recommended handling measures, etc.) to the terminal devices of management personnel (such as mobile phones, computers, etc.) in real time to ensure that management personnel can receive notifications and take emergency response measures immediately to minimize the losses caused by anomalies to production;

[0211] 5. Model Update and Optimization

[0212] Data Accumulation and Labeling:

[0213] Continuously collect data during the production process, accumulate data on newly emerging anomalies and data under normal production conditions, and accurately label the anomaly data, recording detailed information such as anomaly type, occurrence time, handling process, etc.; this data will be used for subsequent model updates and optimizations to improve the model's recognition ability and accuracy for different anomaly situations;

[0214] Regular Model Evaluation and Update Trigger:

[0215] Regularly (such as weekly, monthly) evaluate the performance of the deep autoencoder model, and calculate evaluation metrics such as accuracy, recall rate, F1 value, etc. of the model in detecting anomalies using the newly accumulated data (including normal data and labeled anomaly data); if it is found that the model performance shows a downward trend (such as a decrease in accuracy, an increase in false negative or false positive rates), trigger the model update process;

[0216] Model Retraining and Optimization:

[0217] Merge the newly accumulated data with the original training data, re-partition the training set and validation set (the ratio can be appropriately adjusted according to the data distribution), and retrain the deep autoencoder model using the updated dataset; during the training process, try to adjust the hyperparameters of the model (such as learning rate, hidden layer structure, etc.) or adopt more advanced training algorithms to further optimize the model performance; by continuously updating and optimizing the model, make it adaptable to various changes and new anomaly patterns that may occur in the production process of bamboo chopsticks, and continuously improve the accuracy and reliability of anomaly detection.

[0218] In this embodiment, the device collaboration and optimization module is communicatively connected to the intelligent scheduling module and the anomaly detection module. Using the federated learning algorithm, based on the production scheduling plan generated by the intelligent scheduling module and the anomaly situation feedback by the anomaly detection module, it optimizes the collaborative working mode among the devices in the bamboo chopstick production line. The device collaboration and optimization module includes:

[0219] Inter-device task allocation optimization unit: Under the federated learning architecture, each device is regarded as a node participating in collaboration. Based on the production scheduling plan generated by the intelligent scheduling module, combined with the current real-time status of each device and the anomaly information feedback by the anomaly detection module, an optimization model for inter-device task allocation is constructed. Let the device set be , be the total number of devices, the production task set be , be the number of production tasks. Define the allocation matrix , where the element indicates whether task is allocated to device . The goal is to find the optimal allocation matrix . The optimization process is achieved by iteratively updating the allocation strategy among the device nodes through the federated learning algorithm, that is, each device node adjusts its corresponding value based on its local status and the globally shared information. is the index of the device, is the index of the production task;

[0220] Considering the heterogeneity of the devices, for each device set the task processing capacity vector , which represents its processing efficiency or capacity limit for different types of tasks. When optimizing the task allocation, add a constraint condition to ensure that the total amount of tasks allocated to device does not exceed its processing capacity, that is, , where represents the workload of task , represents the processing capacity limit of device for the task belonging to the task type to ensure the rationality and feasibility of task allocation;

[0221] Device operation parameter collaborative adjustment unit: Each device shares its own operation parameters through the federated learning mechanism to construct a collaborative adjustment model of operation parameters based on multiple devices. Let the operation parameter vector of device be . By analyzing the correlation between the operation parameters of different devices and their impact on the overall production;

[0222] Update the operating parameters of each device under the federated learning framework using the gradient descent optimization algorithm. Let the learning rate be , and in each iteration, the operating parameter update rule for device is , where represents the updated operating parameter vector of device , is the learning rate, represents the gradient of the overall production line performance loss function with respect to the operating parameter vector of device , and is the overall production line performance loss function;

[0223] Further, the specific operation steps of the device collaboration and optimization module are as follows:

[0224] 1. Data acquisition and preprocessing

[0225] Establish communication connections:

[0226] The device collaboration and optimization module uses Ethernet, Wi-Fi, or a specific industrial communication protocol to establish reliable two-way communication links with the intelligent scheduling module and the anomaly detection module to ensure the stability and real-time nature of data transmission;

[0227] Receive scheduling scheme data:

[0228] Obtain detailed production scheduling schemes from the intelligent scheduling module, including the device task assignment list (specifying the specific production tasks undertaken by each device, such as the processes, quantities, and estimated processing times for processing bamboo chopsticks), the set values of device operating parameters (such as the specific values of process parameters such as the rotation speed, temperature, and pressure of the device), and the task priority ranking (distinguishing between urgent tasks and regular tasks);

[0229] Receive anomaly situation data:

[0230] Receive anomaly information sent by the anomaly detection module in real time, covering device failure details (such as device name, number, failure code, occurrence time), raw material shortage warnings (types of shortage raw materials, quantities, estimated impact duration), and production bottleneck reports (name, location of the bottleneck link, comparison of the current processing capacity with the normal level);

[0231] Data cleaning and integration:

[0232] Clean the obtained scheduling scheme data and anomaly situation data to remove invalid, duplicate, and incorrect data records; for example, correct abnormal data values caused by sensor failures or communication interference, and delete duplicate data entries;

[0233] Integrate the cleaned data according to device identification, time sequence, etc. to form a dataset in a unified format for convenient subsequent analysis and processing; for example, associate the task allocation information of a certain device with the set values of the corresponding operating parameters of the device, and mark whether there are abnormal situations and the types of abnormalities;

[0234] 2. Build a federated learning model

[0235] Determine the model architecture:

[0236] According to the characteristics and optimization objectives of the collaborative work of the bamboo chopstick production line equipment, select a suitable federated learning model architecture, such as a federated learning model based on a deep neural network or a federated decision tree model, etc.;

[0237] If a deep neural network is adopted, design the network structure, including an input layer (receiving feature data such as device status, task information, abnormal identification, etc.), a hidden layer (used to extract data features and patterns, multiple layers can be set, and the number of layers and nodes are determined according to data complexity), and an output layer (outputting optimization strategies for the collaborative work of the device, such as task adjustment suggestions, device parameter correction values, etc.);

[0238] Model initialization:

[0239] On each device node (client) participating in federated learning, perform an initialization operation on the selected model; assign the same initial model parameters to each client, such as the initial weights and biases of the neural network, to ensure that the model has a unified starting state globally for subsequent collaborative training and optimization;

[0240] 3. Optimization of task allocation among devices

[0241] Local task allocation model training (client):

[0242] Each device client constructs a local task allocation model based on local data (including its own device status, received task information, and some shared production line data);

[0243] With the goal of minimizing the production cycle and maximizing the equipment utilization rate, considering the real-time status of the equipment (such as the current load of the equipment, remaining available time, operating efficiency, etc.), task characteristics (such as task urgency, processing difficulty, expected processing time, etc.), and the impact of abnormal situations (such as task interruption caused by equipment failure, adjustment of task priority due to raw material shortage), establish a local objective function;

[0244] Adopt the gradient descent algorithm or other optimization algorithms to train the local model according to the local objective function and data, and optimize the model parameters (such as weights and biases in the neural network) so that the model can generate reasonable task allocation suggestions according to the input data;

[0245] Model Parameter Upload and Aggregation (Server Side):

[0246] After a certain number of rounds of local training, each client encrypts and uploads the parameters of its local model (such as the weight matrix and bias vector of a neural network) to the central server or coordinator node;

[0247] The central server uses the Federated Averaging (FedAvg) algorithm or other aggregation algorithms to aggregate the uploaded client parameters; during the aggregation process, weights are assigned to clients based on factors such as their importance in the production line and data quality, and the updated parameters of the global model are calculated; for example, for key production equipment or clients with high data accuracy, their parameters are given higher weights during aggregation;

[0248] Global Model Parameter Download and Update (Client):

[0249] The central server downloads the aggregated global model parameters back to each client;

[0250] The client uses the received global parameters to update its local model, synchronizing the local model with the global model and achieving collaborative optimization of the model on a global scale;

[0251] By iterating the above local training, parameter upload, aggregation, and download and update processes multiple times, the task allocation strategy is continuously optimized, making the task allocation among devices more reasonable and improving the overall production efficiency of the production line;

[0252] 4. Collaborative Adjustment of Equipment Operating Parameters

[0253] Local Operating Parameter Adjustment Model Training (Client):

[0254] Each device client constructs a local equipment operating parameter collaborative adjustment model based on local device operating data (such as real-time device operating parameters, historical operating data, device performance indicators, etc.) and the received abnormal situation information;

[0255] With the goal of minimizing energy consumption and ensuring stable product quality, considering the correlation between equipment operating parameters (such as the mutual influence between equipment speed and temperature, pressure), the current state of the equipment (such as whether there are potential faults, equipment wear degree, etc.), and production task requirements (such as product specifications' limitations on equipment parameters), a local objective function is established;

[0256] Using the gradient descent optimization algorithm, the local model is trained according to the local objective function and data, and the model parameters are adjusted so that the model can generate appropriate equipment operating parameter adjustment schemes based on the input data;

[0257] Parameter Upload, Aggregation, and Update (Similar to Task Allocation Optimization):

[0258] After the client is trained for a certain number of rounds, the gradient information related to the local model running parameters is encrypted and uploaded to the central server;

[0259] The central server decrypts and aggregates the uploaded gradient information (also considering the client weights), calculates the global gradient, and then updates the global model parameters according to the global gradient;

[0260] The central server sends the updated global parameters back to the client, and the client updates the local model accordingly, realizing the collaborative optimization of device operation parameters within the global scope, ensuring the mutual matching of device operation parameters, and improving the overall performance of the production line;

[0261] 5. Exception Handling and Collaborative Adjustment

[0262] Analysis of Abnormal Situations and Generation of Strategies:

[0263] After receiving the abnormal situation data, the device collaboration and optimization module quickly conducts analysis and evaluation; for device failures, determine the failure type, severity, and the scope of its impact on the production process (such as which subsequent processes are affected and which device tasks need to be adjusted); for raw material shortages, evaluate the impact of the shortage degree on the production of different products and the possible production downtime; for production bottlenecks, analyze the reasons for the bottleneck (such as equipment aging, unreasonable processes, insufficient personnel allocation, etc.) and the degree of restriction of the bottleneck on the efficiency of upstream and downstream equipment and the entire production line;

[0264] According to the abnormal analysis results, combined with the current production scheduling plan and device status, formulate targeted collaborative adjustment strategies; for example, in case of device failures, determine which standby or idle devices the tasks of the failed device will be transferred to, and how to adjust the operating parameters of relevant devices to adapt to the new tasks; in case of raw material shortages, adjust the production task priorities, give priority to producing products with low raw material consumption or products that can be supported by the in - stock raw materials, and coordinate with the purchasing department to accelerate raw material procurement; in case of production bottlenecks, take measures such as optimizing the process parameters of bottleneck devices, adding temporary auxiliary devices, or adjusting the task allocation of upstream and downstream devices to alleviate the impact of the bottleneck on production;

[0265] Execution and Monitoring of Collaborative Adjustment:

[0266] Execution of Task Re - allocation: According to the formulated strategy, using the optimized task allocation model, quickly calculate the new task allocation plan for each device in case of abnormal situations; during the task re - allocation process, fully consider the real - time status of devices (such as the remaining processing capacity of available devices, device load balancing), changes in task priorities (urgent tasks first), and production process constraints (ensuring the continuity and rationality of the production process); timely issue the adjusted task allocation plan to each device control system to ensure that the devices can execute tasks according to the new plan;

[0267] Dynamic adjustment execution of equipment parameters: According to abnormal situations and adjustment strategies, through the operation parameter collaborative adjustment model, calculate the adjustment values of the operation parameters of each relevant equipment in real time; for example, when task transfer is caused by equipment failure, appropriately increase its operation parameters (such as rotation speed, power, etc.) according to the actual situation of the equipment receiving the task to speed up the production progress; when there is a shortage of raw materials, reduce the operation parameters of the equipment (such as temperature, output, etc.) to reduce raw material consumption; when there is a bottleneck in the production process, optimize the operation parameters of the bottleneck equipment and its upstream and downstream equipment to achieve a new balance in each link of the production line; send the adjusted operation parameters to the equipment in real time to achieve dynamic optimization of the equipment operation parameters;

[0268] Collaborative adjustment monitoring: After implementing the collaborative adjustment measures, continuously monitor the operation status of the production line, and collect equipment operation data (such as equipment utilization rate, failure rate, energy consumption, etc.), production progress data (completion time of each process, product output, etc.) and product quality data (qualified rate, defective rate, etc.) in real time; by comparing the data before and after the adjustment, evaluate the effect of the collaborative adjustment measures and judge whether the production line has returned to normal operation or achieved the expected optimization goal;

[0269] Feedback and optimization loop:

[0270] Feed back the evaluation results of the collaborative adjustment effect to the intelligent scheduling module and the anomaly detection module to provide reference for subsequent decision-making; if there are still problems in the production line after the adjustment or the expected effect is not achieved, such as low production efficiency, unbalanced equipment load or unstable product quality, further analyze the reasons based on the feedback information, and it may be necessary to optimize the adjustment strategy, federated learning model parameters or algorithms;

[0271] At the same time, summarize the experiences and lessons in the process of anomaly handling and collaborative adjustment, store the successful cases and optimization strategies in the knowledge base, provide a reference basis for the handling of future similar abnormal situations, continuously improve the equipment collaboration and optimization mechanism, and improve the ability of the production line to handle abnormal situations and the overall operation efficiency.

[0272] In this embodiment, the decision support module is communicatively connected to the intelligent scheduling module, the anomaly detection module, and the equipment collaboration and optimization module. Based on the data obtained from each module, through big data analysis technology, a visual interface for managers to view is generated, and an intelligent analysis report covering production efficiency, equipment utilization rate, and raw material consumption is generated regularly. The decision support module includes:

[0273] Real-time monitoring visualization unit: Build a visual interface, and on the interface, intuitively display the detailed status information of the bamboo chopstick production line in the form of charts, graphs, and data lists, covering equipment operation status, production progress, the currently effective production scheduling plan, and equipment health status warning information;

[0274] Intelligent Report Generation Unit: According to the preset time period, comprehensively analyze the historical production data and real-time data of the bamboo chopstick production line using big data analysis technology, and generate an intelligent analysis report covering production efficiency, equipment utilization rate, and raw material consumption;

[0275] In the intelligent analysis report, in addition to presenting the actual data and change trends of each indicator, data mining algorithms are also used to mine potential problems, optimization directions, and improvement suggestions existing in the production process;

[0276] Furthermore, the specific operation steps of the decision support module are as follows:

[0277] 1. Data collection and integration

[0278] Establish communication connection and data acquisition:

[0279] The decision support module establishes real-time communication connections with the intelligent scheduling module, anomaly detection module, and equipment coordination and optimization module through stable network communication technologies (such as Ethernet, Wi-Fi, etc.) to ensure that data can be transmitted in a timely and accurate manner;

[0280] Obtain production scheduling-related data from the intelligent scheduling module, including production task allocation plans, equipment operation plans, task priority information, etc.; obtain anomaly event data from the anomaly detection module, such as equipment failure records, raw material shortage warnings, production bottleneck information, etc., as well as detailed descriptions of the time, location, and severity of anomalies; obtain equipment operation status data (such as real-time equipment operation parameters, equipment utilization rate, equipment failure repair status, etc.) and the execution result data of equipment coordination work adjustment strategies from the equipment coordination and optimization module;

[0281] Data cleaning and transformation:

[0282] Clean the collected data to remove noise data, error data, and duplicate data; for example, verify and correct device operation parameter values that are clearly outside the normal range (such as too high temperature or too low pressure), and delete duplicate data records caused by communication failures;

[0283] Unify the data in different formats and units into a format suitable for analysis and processing, such as unifying the time format into the standard date and time format, unifying the units of equipment operation parameters into international standard units, and converting production volume data into the same counting unit (such as pieces, tons, etc.);

[0284] Data integration and association:

[0285] Integrate data from different modules according to the logical relationships in the production process to establish associations between data. For example, associate production task assignment information with the operating data of the equipment performing the task, and correspond abnormal events to the affected production links and equipment, so as to comprehensively understand the interrelationships between various factors in the production process and provide a basis for subsequent analysis and visualization.

[0286] 2. Construction and Update of Real-time Monitoring Visualization Interface

[0287] Visualization Interface Design and Layout:

[0288] Design the layout and display content of the visualization interface according to the needs of management personnel and the key points of production management. The interface can be divided into multiple areas, such as the equipment operating status monitoring area, production progress display area, abnormal warning area, production index summary area, etc.

[0289] In the equipment operating status monitoring area, graphically display (such as through dashboards, line charts, bar charts, etc.) the key operating parameters (such as temperature, rotational speed, pressure, etc.), operating status (running, stopped, faulty, etc.) of each equipment, and the trend of equipment utilization rate in real time. In the production progress display area, intuitively show the completion progress, estimated completion time of each production task, and the comparison with the planned progress through forms such as progress bars and Gantt charts. The abnormal warning area uses eye-catching colors (such as red for serious abnormalities and yellow for minor abnormalities) and flashing icons to prompt the current abnormal events and provide links to abnormal details. The production index summary area displays key indicators such as current production efficiency, equipment utilization rate, and raw material consumption in numerical form.

[0290] Data Real-time Update and Push:

[0291] Utilize data push technology or a timed query mechanism to obtain the latest production data in real time and update it on the visualization interface. Ensure that management personnel can timely understand the real-time status of the production line and make accurate decisions. For example, update the equipment operating parameters and production progress data every few seconds. When an abnormal event occurs, immediately display relevant information in the abnormal warning area and push notifications to management personnel.

[0292] Interactive Function Design:

[0293] Add interactive functions to the visualization interface to facilitate managers' in-depth understanding of production situations. For example, set click events so that when managers click on the equipment icon or the production task progress bar, a detailed information window pops up, displaying the detailed data, historical records, and relevant operation logs of the equipment or task. Provide zooming and panning functions to view production data for different time periods or different equipment groups. Support data filtering and sorting functions, allowing managers to view data under specific conditions according to their needs, such as viewing the operation data of specific equipment within a certain time period or sorting production tasks by production efficiency.

[0294] 3. Intelligent Analysis Report Generation

[0295] Data Statistics and Analysis:

[0296] Conduct a comprehensive statistical analysis of the integrated production data according to a predetermined time cycle (such as daily, weekly, monthly). Calculate production efficiency, which is the ratio of the output of qualified products per unit time to the theoretical maximum output, analyze its changing trends in different time periods, and identify factors affecting production efficiency, such as equipment failures, production process adjustments, and personnel operation levels. Statistically analyze equipment utilization rate, calculate the percentage of the actual operating time of each equipment to the planned operating time, evaluate the usage efficiency and idle situation of the equipment, and analyze the relationship between equipment utilization rate and production task allocation. Analyze the raw material consumption situation, calculate the raw material consumption per unit product, compare the raw material consumption differences between different products and different production batches, and identify the reasons for excessive or abnormal fluctuations in raw material consumption, such as changes in raw material quality and unreasonable production processes.

[0297] Problem Mining and Optimization Suggestions:

[0298] Use data mining algorithms (such as association rule mining, clustering analysis, decision tree algorithms, etc.) to deeply mine production data and discover potential production problems and optimization opportunities. For example, through association rule mining, discover the association relationships between certain equipment failure modes and specific production tasks or raw material batches, providing a basis for equipment maintenance and production plan adjustment. Use clustering analysis to classify similar working conditions or product quality problems in the production process, identify the commonalities and differences of the problems, and take targeted improvement measures. Use decision tree algorithms to analyze the impact of different production decisions on production indicators and provide a reference for optimizing production decisions.

[0299] Generate corresponding optimization suggestions based on the data analysis results and the discovered problems. For example, if it is found that the utilization rate of a certain equipment is low for a long time, it is recommended to adjust the production task allocation or consider equipment upgrade and transformation. If the raw material consumption is too high, suggestions for optimizing the production process, improving raw material quality control, or finding alternative raw materials are put forward. For frequent abnormal events in the production process, measures to strengthen the equipment maintenance plan, optimize the abnormal warning threshold, or improve the production process are formulated.

[0300] Report generation and distribution:

[0301] Integrate the statistical analysis results, problem discovery, and optimization suggestions to generate an intelligent analysis report. The report content includes written descriptions, data charts, data tables, as well as the specific steps and implementation plans for problem solutions and suggestions;

[0302] Regularly distribute the intelligent analysis report to relevant management personnel via email, enterprise internal system message push, or generating downloadable files, etc., to ensure that management personnel can obtain the report content in a timely manner and make decisions and optimize management based on the information provided in the report; at the same time, store the report in the enterprise data warehouse or document management system for convenient historical data query and comparative analysis, providing data support for the continuous improvement of enterprise production management.

[0303] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A real-time intelligent scheduling and management system for bamboo chopstick production line data, characterized by: include: Data acquisition module: It is set up on the bamboo chopstick production line and connected with sensors, PLC controllers and production equipment interfaces to collect real-time equipment operation status data, production progress data, raw material consumption data and output statistics during the bamboo chopstick production process; Intelligent scheduling module: It communicates with the data acquisition module and uses the federated learning algorithm based on the collected equipment operation status data, production progress data, raw material consumption data and output statistics. It also considers the real-time status of the equipment, current production needs and task priorities, and automatically generates a production scheduling plan for the bamboo chopstick production line through intelligent calculations. Abnormality detection module: communicates with the data acquisition module and uses deep autoencoder technology to analyze the collected data in real time to monitor abnormal conditions in the production process of bamboo chopsticks, including equipment failure, insufficient raw materials, and bottlenecks in the production process; Equipment coordination and optimization module: It communicates with the intelligent scheduling module and the anomaly detection module, and uses the federated learning algorithm to optimize the collaborative working mode between the various devices in the bamboo chopstick production line based on the production scheduling plan generated by the intelligent scheduling module and the anomaly feedback from the anomaly detection module; Decision support module: communicates with the intelligent scheduling module, anomaly detection module, and equipment coordination and optimization module. Based on the data obtained from each module, it generates a visual interface for management personnel to view through big data analysis technology, and regularly generates intelligent analysis reports covering production efficiency, equipment utilization, and raw material consumption; The intelligent scheduling module includes: Multi-objective collaborative optimization unit: In the application of the federated learning algorithm, multiple equipment nodes or links involved in the scheduling of the bamboo chopstick production line are set as clients of the federated learning, and the production efficiency is set as , equipment utilization is set to , energy consumption is set to , Raw material waste is set to As a joint optimization target; each client Based on locally collected data Constructing a local objective function ,in, For Clients The decision variable vector, the overall multi-objective optimization objective function is expressed as: ,in, is the number of clients, For Clients The weights of the models are exchanged and updated among the clients through the federated learning mechanism to collaboratively optimize the various objectives. Client The local objective function The construction method is as follows: Assume that the product output per unit time is , the equipment running time is ,but , Represents the client production efficiency, Comprehensive consideration It is constructed with other local related factors and constraints according to preset rules and mathematical relationships to accurately reflect the local contribution to the overall multi-objective optimization; Adaptive adjustment unit: collects production-related data from each client in real time and aggregates them into a real-time status data set for the production line , and set a predefined set of emergency adjustment rules ; When abnormal situations occur, such as urgent order insertion and sudden equipment failure, the global coordination mechanism based on federated learning is used. and , by updating the scheduling policy parameters of each client To adjust the production scheduling plan, the update process is expressed as ,in, is the adaptive adjustment function based on federated learning, is the updated scheduling strategy parameter; For different types of abnormal situations, The corresponding specific adjustment strategy details are set in the system. In case of equipment failure, the system determines how to reallocate production tasks to other normal equipment according to the link where the faulty equipment is located and the remaining available equipment capacity. Through the coordination of federated learning, the adjustment strategy is sent to the corresponding client for execution. Task priority determination unit: Based on the urgency of the delivery date of bamboo chopstick orders , Product type importance level and the adequacy of raw material supply To determine the priority of production tasks; each client Calculate the local priority score based on local relevant information , ,in, For Clients correspond , , The weight coefficient of Global production task priority Through the aggregation mechanism of federated learning, the Get, using weighted average method ,in, For Clients The weight in the global priority calculation.

2. According to claim 1, a real-time intelligent scheduling and management system for bamboo chopstick production line data is characterized in that: The anomaly detection module comprises: Anomaly recognition and positioning unit: Use deep autoencoders to process the collected data and convert the input data Mapped to hidden layer representation through encoder , and then reconstructed by the decoder to get the output , calculate the reconstruction error , set a reasonable threshold ,when When determining that the production link corresponding to the data has abnormal conditions; Analyzing Hidden Layer Representations Based on the characteristic distribution of each device and each production link under normal conditions, the specific abnormal device or production link can be located; Abnormal warning and feedback unit: Once an abnormal situation is detected, an early warning signal containing the abnormality type, location, and severity is immediately generated and sent to the intelligent scheduling module and the equipment coordination and optimization module through the communication link; Different levels of early warning prompts are set according to the severity of the abnormality. For minor abnormalities, prompt information is only displayed on the visual interface. For abnormalities that seriously affect production, sound and light alarms are triggered at the same time, and detailed abnormality information is pushed to the manager's terminal device in real time.

3. The real-time intelligent scheduling and management system for bamboo chopstick production line data according to claim 1 is characterized by: The equipment coordination and optimization module includes: Inter-device task allocation optimization unit: Under the federated learning architecture, each device is regarded as a node participating in the collaboration. According to the production scheduling plan generated by the intelligent scheduling module, combined with the current real-time status of each device and the abnormal information fed back by the abnormal detection module, an optimization model for task allocation between devices is constructed; let the device set be , is the total number of equipment, and the production task set is , is the number of production tasks, defining the allocation matrix , where the element Indicates the task Is it assigned to the device? , the goal is to find the optimal allocation matrix The optimization process is achieved by iteratively updating the allocation strategy among each device node through the federated learning algorithm, that is, each device node adjusts its corresponding value, is the index of the device, is the index of the production task; Considering the heterogeneity of devices, Set the task processing capability vector , indicating its processing efficiency or capacity upper limit for different types of tasks. When optimizing task allocation, adding constraints ensures that the tasks are allocated to the device. The total amount of tasks does not exceed its processing capacity, that is ,in, Indicates the task The workload, Indicates the device To the task Type of task The processing capacity limit of the system is determined to ensure the rationality and feasibility of task allocation; Equipment operation parameter collaborative adjustment unit: Each device shares its own operation parameters through the federated learning mechanism, and builds an operation parameter collaborative adjustment model based on multi-device joint; The operating parameter vector is , by analyzing the correlation between different equipment operating parameters and their impact on overall production; The gradient descent optimization algorithm is used to update the operating parameters of each device under the federated learning framework. The learning rate is set to , in each iteration, the device The operating parameter update rule is ,in, Indicates the device The updated operating parameter vector, is the learning rate, Represents the overall production line performance loss function About the device Run parameter vector The gradient of is the overall production line performance loss function.

4. The real-time intelligent scheduling and management system for bamboo chopstick production line data according to claim 1, characterized in that: The decision support module includes: Real-time monitoring visualization unit: Build a visualization interface to display detailed status information of the bamboo chopstick production line in real time with intuitive charts, covering equipment operation status, production progress, currently effective production scheduling plan and equipment health status warning information; Intelligent report generation unit: According to the preset time period, the unit uses big data analysis technology to conduct a comprehensive analysis of the historical production data and real-time data of the bamboo chopstick production line, and generates an intelligent analysis report covering production efficiency, equipment utilization, and raw material consumption; In the intelligent analysis report, in addition to presenting the actual data and changing trends of each indicator, data mining algorithms are also used to explore potential problems, optimization directions and improvement suggestions in the production process.

Images