Production, supply and marketing integrated automatic management system for ring network equipment and data processing method

Abstract

The invention relates to the technical field of ring network equipment management, and discloses a production, supply and marketing integrated automatic management system for ring network equipment and a data processing method.The production, supply and marketing integrated automatic management system for the ring network equipment comprises a data collection module used for collecting data related to production, purchase, inventory and sales of the ring network equipment in real time; the data preprocessing module is used for performing cleaning, standardization, missing value filling and abnormal value detection on the acquired data to ensure the data quality; the reinforcement learning module is used for dynamically optimizing a production plan, a purchase plan and inventory management by utilizing a reinforcement learning algorithm based on the collected real-time data and historical data; and a multi-objective optimization module. According to the system, strategy adjustment is carried out through reinforcement learning, target conflicts are solved through multi-target optimization, and a feedback adjustment mechanism is provided through a nonlinear dynamic model. And through real-time data updating and report generation, high-efficiency collaboration of links of production, purchasing and inventory is ensured, and the resource allocation and operation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ring network equipment management, and in particular to an integrated production, supply and marketing automated management system and a data processing method for ring network equipment. Background Art

[0002] With the complexity of modern manufacturing and supply chain management, traditional production, procurement, and inventory management systems have exposed many limitations. Especially in the face of rapidly changing market demands and production conditions, existing technologies are often unable to adjust and optimize decisions in real time, resulting in low system efficiency, uneven resource allocation, and excessive operating costs. Traditional technologies mainly rely on static rules and linear models, and fail to fully consider the dynamic characteristics of production and supply chain systems, making it difficult to meet the needs of modern enterprises for rapid response and flexible adjustment.

[0003] One of the main problems of existing technologies is the lack of effective support for multi-objective optimization. In traditional production, procurement, and inventory management, most optimization models focus only on a single objective (e.g., cost minimization or inventory level control), while ignoring other important objectives that may affect system performance, such as capacity utilization and supply chain response speed. This one-sided optimization approach not only fails to find the best balance between multiple objectives, but may also lead to irrational resource allocation, increasing operational risks and enterprise costs.

[0004] In addition, many existing systems are modeled based on linear assumptions and cannot accurately simulate the complex, nonlinear relationships between production, inventory, and procurement. Traditional linear optimization models fail to take into account the mutual influence and feedback effects of various links in the actual system, making it impossible for companies to respond quickly to production plans and inventory management in a rapidly changing market environment, resulting in a series of problems such as production delays and inventory backlogs. At the same time, traditional systems often have information islands and lack comprehensive data integration and sharing, which leads to the inability of various links to work together, affecting the efficiency and accuracy of overall decision-making. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an integrated production, supply and marketing automated management system and a data processing method for ring network equipment. The existing system is based on linear assumption modeling and cannot effectively simulate the nonlinear relationship between production, inventory and procurement, resulting in an inability to quickly respond to market changes, and causing problems such as production delays and inventory backlogs. In addition, the system lacks data integration and collaboration, which affects decision-making efficiency.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an integrated production, supply and marketing automated management system for ring network equipment, comprising: Data collection module, used to collect data related to production, procurement, inventory and sales of ring network equipment in real time; Data preprocessing module, used to clean, standardize, fill missing values ​​and detect outliers on the collected data to ensure data quality; Reinforcement learning module, based on the collected real-time data and historical data, uses reinforcement learning algorithms to dynamically optimize production plans, procurement plans and inventory management; Multi-objective optimization module optimizes multiple objectives such as production cost, inventory holding cost, and capacity utilization to obtain the best decision-making solution; Nonlinear dynamics module, which models the dynamic relationship between production, procurement and inventory based on nonlinear dynamics model to ensure coordinated operation of each link; The decision support module generates production plans, procurement plans and inventory management strategies based on the output results of the reinforcement learning module and the multi-objective optimization module, and makes adjustments through real-time feedback.

[0007] Preferably, the data acquisition module includes an Internet of Things device and an API interface, which is used to collect real-time data from the production line, warehouse and sales system of the ring network equipment, and the data includes production progress, production capacity, raw material consumption, market demand, customer orders, and inventory consumption rate.

[0008] Preferably, the reinforcement learning module trains the intelligent agent based on real-time data and historical data through Q learning or deep Q network algorithm, so that the intelligent agent can adaptively adjust the production plan, procurement plan and inventory management strategy according to environmental changes, and dynamically adjust the reward function in the reinforcement learning process through a real-time feedback mechanism to ensure real-time optimization of the intelligent agent strategy.

[0009] Preferably, the multi-objective optimization module uses a weighted sum method to weight multiple objectives and convert them into a single-objective optimization problem for solution. The objectives include production cost, inventory cost, and capacity utilization, wherein the weight of the objective function is customized by the user to achieve a trade-off between different objectives.

[0010] Preferably, the nonlinear dynamics module is dynamically adjusted through an automated management system, using the Lotka-Volterra model to simulate the interaction between production, procurement and inventory links, and adjust the control strategy of each link based on real-time data. The model is used to dynamically optimize feedback decisions to ensure continuous coordination and efficient operation of each link.

[0011] Preferably, the decision support module includes a real-time data update module, which is used to adjust and optimize the production plan, procurement plan and inventory management strategy in real time according to the actual changes in production, procurement, sales and inventory, and continuously adjust the optimization plan according to the feedback mechanism to form a closed loop of continuous improvement and optimization.

[0012] Preferably, the data collection module is connected to a centralized data processing platform, which aggregates and classifies the collected production progress, inventory status, raw material consumption, and customer order data, and generates a real-time decision support report through preset report generation rules.

[0013] Preferably, the reinforcement learning module receives real-time data on market demand changes, supplier delivery dates and production capacity.

[0014] Preferably, the multi-objective optimization module sequentially performs the following steps based on the input data of production cost, inventory holding cost, and capacity utilization rate collected in real time: Accepts user-defined objective weights and transforms multiple objectives into a single-objective optimization problem; Construct objective functions based on real-time and historical data and determine constraints; Obtain the global optimal solution through iterative optimization algorithm; Output optimized production plans, procurement plans and inventory management strategies, and adjust relevant parameters in real time to adapt to environmental changes.

[0015] The present invention also provides a production, supply and marketing integrated data processing method for ring network equipment, comprising the following steps: Collect data related to ring network equipment production, procurement, inventory and sales; Clean and standardize the collected data to ensure data consistency and quality; Dynamically optimize production plans, procurement plans, and inventory management based on reinforcement learning algorithms, and the agent adjusts its decision-making strategy based on historical data and real-time feedback; Use a multi-objective optimization algorithm to optimize multiple objectives such as production cost, inventory cost and capacity utilization to obtain the best decision-making solution, where the objective weights of the multi-objective optimization are defined by the user; Use nonlinear dynamic models to model the dynamic relationship between production, procurement and inventory, and adjust decisions based on feedback; Generate production plans, procurement plans and inventory management strategies based on optimization results, and make decision adjustments through real-time data update mechanism; The optimization plan is continuously adjusted through the feedback mechanism to ensure that all links are always in efficient coordination and optimal state.

[0016] The present invention provides an integrated production, supply and marketing automated management system and data processing method for ring network equipment. It has the following beneficial effects: 1. The present invention adopts a technical solution that combines reinforcement learning algorithm with real-time data feedback to achieve dynamic optimization of production planning, procurement planning and inventory management. Compared with the static optimization method in the prior art, the present invention can adjust the decision in real time according to system feedback, solving the problem that traditional optimization solutions cannot adapt to market changes and fluctuations in the production environment.

[0017] 2. The present invention uses a multi-objective optimization algorithm to comprehensively consider multiple objectives such as production cost, inventory cost and capacity utilization, and achieves the global optimal decision of multi-objective optimization. Compared with the technical solution of single-objective optimization in the prior art, it solves the problem of conflict between objectives in various links and ensures the efficient allocation and utilization of resources.

[0018] 3. The present invention uses a nonlinear dynamic model to model the dynamic relationship between production, procurement and inventory, and realizes real-time adjustment of decisions through a feedback mechanism. Compared with the prior art method of using only linear models, the present invention can more accurately simulate and predict the complex interactions between various links, solving the problem that the linear assumption cannot truly reflect the complexity of the system.

[0019] 4. The present invention realizes intelligent collaborative management of production, procurement and inventory by integrating technologies such as data cleaning, standardization, optimization algorithms and nonlinear dynamic models. Compared with the existing solutions that process data and decisions in isolation at each link, the present invention solves the problem that data at each link cannot be effectively circulated and work collaboratively through systematic integration, greatly improving management efficiency and decision-making accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a system architecture diagram of the present invention; Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Please see attached Figure 1 The embodiment of the present invention provides an integrated production, supply and marketing automated management system for ring network equipment, including: Data collection module, used to collect data related to production, procurement, inventory and sales of ring network equipment in real time; In this embodiment, the data acquisition module is connected to each link of the ring network equipment through the Internet of Things (IoT) devices, sensors and API interfaces to obtain multi-dimensional data such as production progress, raw material consumption, inventory status, market demand, customer orders, supplier delivery time, etc. in real time. These collected data will be summarized and transmitted to the system's data platform for use by subsequent modules (such as data preprocessing, reinforcement learning, multi-objective optimization, nonlinear dynamic models, etc.). Each type of data has a specific collection method and standard to ensure accuracy and consistency.

[0023] Internet of Things Devices (IoT): Sensors: used to monitor the status of the production line (such as temperature, humidity, pressure, vibration, etc.) and the operation of the equipment in real time. By monitoring the key parameters of the production line in real time, it is possible to ensure that no abnormalities occur during the production process, thereby dynamically adjusting the production plan.

[0024] RFID and barcode scanners: Used to track the consumption of raw materials and the production, storage and distribution paths of products in real time. Each raw material or finished product has a unique identifier, and inventory data can be updated in real time by scanning these tags.

[0025] Cameras and video surveillance: used for visual inspection of production lines, such as quality control and defect detection during the production process. The image data collected by the camera can also be combined with production data for analysis through image recognition technology.

[0026] API interface: The data collection module interacts with existing enterprise management systems (such as ERP, MES, WMS, etc.) through standardized API interfaces. These interfaces support multiple data formats (such as JSON, XML) and data transmission protocols (such as RESTful API, SOAP), ensuring that the system can obtain information such as purchase orders, inventory status, sales data, etc. from the outside.

[0027] Specifically, each time the API interface is updated, the system will regularly pull production, procurement, sales and inventory data from the external system, aggregate and process it, and then transmit it to the data platform.

[0028] Production data: Production progress: The work progress of the equipment and personnel on the production line is updated in real time. For example, data such as production speed, production stage, and product output in each production cycle obtained by sensors.

[0029] Raw material consumption: The amount of raw materials consumed in each production link, including consumption time, consumption type, etc. Real-time feedback from sensors and RFID tags can help the system accurately record the raw material consumption of each production batch.

[0030] Production capacity and equipment status: Real-time recording of the operating status, production capacity, and production speed of production equipment. Sensors on production equipment can detect whether the equipment is idle, running, or faulty, helping to adjust production plans in real time.

[0031] Purchasing data: Supplier delivery time: Record the delivery time obtained from suppliers and automatically update the purchase order status, including delays, deviations from the agreed time, etc.

[0032] Purchase order data: Order data extracted by the system from the purchasing management system, including purchase quantity, unit price, purchase category, etc. This information is connected to the supplier's system through the API interface to synchronize the status of the purchase order in real time.

[0033] Sales data: Market demand: Based on market forecasts and historical sales data, the market demand data is dynamically updated. The system will adjust production plans and inventory management according to changes in market demand.

[0034] Customer order data: Orders submitted by customers through different sales channels (such as online, offline, telephone, etc.), including order amount, delivery time, product specifications, customer-specific requirements, etc.

[0035] Inventory data: Inventory: Inventory recorded in real time through RFID tags and barcode scanners. Including raw material inventory, semi-finished product inventory and finished product inventory. Inventory is reflected in the system in real time to ensure that inventory data is always accurate.

[0036] Inventory consumption rate: Based on sales data and historical consumption data, calculate the inventory consumption rate within a certain period of time. This data helps predict future purchasing needs.

[0037] Inventory warning: Based on inventory levels and demand fluctuations, the system will automatically generate inventory warnings to remind managers to replenish inventory or adjust production plans in a timely manner.

[0038] Data compression and transmission: Due to the large amount of collected data, the system will compress the data to reduce bandwidth consumption and storage requirements. During data transmission, encryption protocols (such as SSL and TLS) are used to ensure data security and avoid data leakage or tampering during transmission.

[0039] In some embodiments, the collected sensor data will be transmitted in batches at certain time intervals to ensure that the system can record and process the data of all devices in a timely manner.

[0040] Data storage and update: The collected data will be stored in the cloud or local server and managed through the database. Real-time data will be stored in the form of time series for subsequent analysis and modeling. The system can update the database in real time as needed to ensure the integrity and timeliness of the data.

[0041] Specifically, the data storage system will mainly store structured data, and some unstructured data (such as pictures and video data) will be stored in a distributed storage system.

[0042] The collected data will be transmitted to other functional modules after data preprocessing, such as reinforcement learning module, multi-objective optimization module, nonlinear dynamics model, etc. These modules will use the real-time collected data to adjust production plans, procurement plans and inventory strategies.

[0043] The reinforcement learning module takes the collected production data, procurement data and inventory data as input, and uses Q learning or deep Q network algorithm to make dynamic adjustments to optimize system decisions.

[0044] The multi-objective optimization module will use the collected sales data, inventory data and production data to perform multi-objective optimization and balance multiple objectives such as production cost, inventory cost and capacity utilization.

[0045] The nonlinear dynamics module will simulate the dynamic changes in production, inventory, procurement and other links based on the collected data, and predict possible future trends to adjust production and inventory strategies.

[0046] During the data collection process, multiple formula calculations are involved. The following are some key formulas and parameter definitions: Production progress calculation formula: Production progress In time The time is calculated as follows: ; in: For time Production progress at the time; is the current production volume; The planned production volume.

[0047] Inventory consumption rate calculation formula: Inventory consumption rate In time The calculation formula of time is: ; in: For time Inventory consumption rate at the time of is the inventory consumed during the time period; is the total inventory during this time period.

[0048] Procurement demand forecast formula: Procurement demand forecast The calculation is based on sales data and inventory consumption rate, the formula is as follows: ; in: For time Procurement needs at the time; Forecasting market demand; is the inventory consumption rate.

[0049] Data preprocessing module, used to clean, standardize, fill missing values ​​and detect outliers on the collected data to ensure data quality; In this embodiment, the data preprocessing module is closely connected with the data acquisition module. After the data acquisition module transmits the raw data to the data preprocessing module, the latter is responsible for further processing and integration of the data to ensure the consistency, integrity and reliability of the data. The output results of the data preprocessing will be passed to the reinforcement learning module, the multi-objective optimization module and the nonlinear dynamics model module to complete the subsequent decision support and optimization.

[0050] Data preprocessing includes multiple tasks, including data cleaning, missing value processing, data standardization, outlier detection, data formatting and conversion, etc. Each step is crucial to ensure that the system can obtain high-quality input when processing complex data.

[0051] Data cleaning: The task of data cleaning is to deal with invalid or inconsistent data caused by sensor failure, data collection errors or other system failures. Common cleaning methods include removing null values, filling missing data, deleting duplicate data, etc.

[0052] Generally, for data with empty values ​​or errors, the system will repair them through the following methods: Use the mean or median of historical data to fill missing values; use linear interpolation of previous and subsequent time points to fill missing data; use algorithms to infer missing values. If the inference error is large, mark it as data that requires manual correction.

[0053] Missing value processing: During the data collection process, missing values ​​may appear, especially when sensors or devices do not work properly. The existence of missing values ​​will directly affect the accuracy of data analysis, so effective measures need to be taken to fill these missing data.

[0054] As an option, in this embodiment, time series interpolation is used to fill missing values, especially in production and sales data, where data loss often occurs. According to the trend of historical data, missing data is filled by interpolation method. For example, linear interpolation is performed using the values ​​of two moments before and after.

[0055] Specifically, if the missing data occurs at a certain point in time , linear interpolation can be performed by the following formula: ; in, To fill in the missing values, is the valid data point before the missing value is a valid data point after the missing value.

[0056] Data standardization: When processing data from different sources, the scales and units of various data may be different. In order to enable different types of data to be calculated in the same model, the data must be standardized. Standardization helps eliminate the imbalance problem caused by different data dimensions.

[0057] Specifically, the Z-Score standardization method is used in this embodiment to convert the data into a standard normal distribution with a mean of 0 and a variance of 1. This processing method helps to improve the convergence speed and stability of the model.

[0058] ; in: is the original value of the data; is the mean of the data; is the standard deviation of the data.

[0059] Outlier detection and processing: Outlier detection is used to identify points in the data that deviate from the normal range. Outliers are usually caused by equipment failure, sensor failure or human error. If these abnormal data are not processed, it will affect subsequent analysis and decision-making.

[0060] In one possible implementation, this embodiment uses the interquartile range method (IQR) to detect outliers and calculate the upper quartile of the data. and lower quartile , and then use the IQR calculation formula to determine outliers: ; If the data point Less than IQR or greater IQR, it is considered an outlier.

[0061] Data formatting and conversion: Data formatting refers to converting data into a standard format to ensure that the system can correctly identify and process it. Usually, data needs to be converted into a unified date and time format, numerical data format, etc. Specifically, if the data comes from different systems (for example, production data comes from the MES system, and inventory data comes from the WMS system), different date formats and timestamp formats need to be converted to ensure that they can be processed uniformly on the same platform.

[0062] The data preprocessing module plays a vital role in the system. Its main task is to provide high-quality, cleaned data for subsequent reinforcement learning, multi-objective optimization, nonlinear dynamics model and other modules. The preprocessed data will be input into these modules to help intelligent decision-making and optimization.

[0063] In the reinforcement learning module, the preprocessed data will be used as the input of the intelligent agent to help the intelligent agent optimize its strategy. The standardized production data, inventory data and procurement data will be used as the state input of the environment. After the reward mechanism feedback, the intelligent agent will learn and optimize the decision.

[0064] In the multi-objective optimization module, processed data (such as inventory consumption rate, production progress, market demand, etc.) will be used to build an optimization model to help optimize decision-making objectives such as production plans and procurement plans.

[0065] In the nonlinear dynamic model, the preprocessed data will be used as input to predict the mutual impact of production and procurement links, and to help with production scheduling, inventory optimization, etc.

[0066] The following are some key formulas and parameter definitions involved in the data preprocessing process: Data standardization formula: Data standardization uses the Z-Score method: ; in: is the original data; is the mean of the data set; is the standard deviation of the data set.

[0067] Missing value interpolation formula: For filling missing values, use the linear interpolation formula: ; in: is the missing value after filling; is the data before the missing value; The data after the missing value.

[0068] Outlier detection formula: Use IQR method to detect outliers: ; in: and are the first and third quartiles of the data set, respectively; is the interquartile range, .

[0069] Reinforcement learning module, based on the collected real-time data and historical data, uses reinforcement learning algorithms to dynamically optimize production plans, procurement plans and inventory management; In this embodiment, the specific implementation of the reinforcement learning module is based on Q-learning and deep Q network (DQN) algorithms. By inputting real-time data from production, procurement, inventory and other links, the intelligent agent can continuously optimize decision-making strategies such as production scheduling, inventory management and procurement planning through trial and error. The reinforcement learning module continuously optimizes decisions through exploration-utilization balance, thereby improving the overall response speed and efficiency of the system.

[0070] Input data: The input data of the reinforcement learning module comes from the aforementioned data acquisition module and data preprocessing module to ensure that its input data is of high quality and consistency. The input data includes but is not limited to: Production data: including production capacity, production progress, equipment status, raw material consumption, etc.; Procurement data: such as supplier delivery date, purchase order, procurement cost, etc.; Inventory data: such as current inventory, inventory consumption rate, inventory replenishment cycle, etc.; Sales data: such as market demand, customer orders, sales progress, etc.

[0071] These data provide real-time status to the reinforcement learning module in the form of time series, reflecting the specific conditions of current production, inventory, procurement, etc. In this way, the reinforcement learning module can make corresponding decisions based on the real-time changing environmental status.

[0072] Output decision: The output of the reinforcement learning module is the best decision based on the current state. These decisions include: Production scheduling decision: Determine the production quantity and production plan based on market demand, production capacity, inventory level and other data; Inventory management decisions: Determine inventory replenishment and adjustment strategies based on inventory consumption rate, inventory warning, sales forecast, etc.; Procurement planning decisions: Optimize procurement strategies, determine order quantities and supplier selection based on supplier delivery dates, procurement costs, market demand, etc.

[0073] Each decision output is based on the assessment of environmental status and optimization of strategies to minimize costs, maximize resources and respond quickly to demand.

[0074] This embodiment uses Q-learning and deep Q network (DQN) algorithms to implement the reinforcement learning module. Q-learning is a reinforcement learning algorithm based on value function. It finds the optimal strategy by updating the Q value, while the deep Q network (DQN) uses a neural network to approximate the Q value function, which is suitable for solving large-scale complex problems.

[0075] Q-learning algorithm: In Q-learning, the agent Evaluation, choose an action , and based on the immediate reward To update the Q value. The core update formula of Q-learning is: ; in, For the status Next select action Q value; is the learning rate, which controls the step size of Q value update; For the status Next action Immediate rewards after is the discount factor, controlling the impact of future rewards; For the next state The maximum Q value of all possible actions.

[0076] In some embodiments, through the Q-learning algorithm, the agent can gradually find the best decision-making strategy through multiple trials and error adjustments. For example, in inventory management, the agent will learn how to make appropriate inventory replenishment decisions based on market demand and inventory levels, thereby reducing inventory backlogs and out-of-stock phenomena.

[0077] Deep Q-Network (DQN): In some complex scenarios, the state space is large, and traditional Q-learning may not be able to obtain effective learning results. In this case, the Deep Q-Network (DQN) can better solve such problems. DQN uses a deep neural network to approximate the Q-value function and further optimize the decision-making process.

[0078] Specifically, the Q-value function of DQN is expressed as: ; in: It is the parameter of the neural network, which is used to train and adjust the approximation of Q value; is the Q value output by the neural network.

[0079] DQN updates network parameters by minimizing the loss function: ; in: are the parameters of the target network; is the parameter of the current network; the loss function measures the difference between the predicted Q value of the current network and the target Q value, and is optimized by the back propagation algorithm.

[0080] In practical applications, the DQN algorithm can adapt to complex production scheduling problems and help the system make fast and effective decisions among a large number of possible production and inventory states.

[0081] The reward mechanism is an important part of reinforcement learning. It guides the learning process by giving feedback to the agent. The goal of reward function design is to guide the agent to make optimal decisions by minimizing costs, improving production efficiency, maintaining inventory balance and other indicators.

[0082] Reward function design: The reward function takes into account the optimization objectives of production, inventory, procurement, etc. Specifically, the reward function usually includes the following important aspects: Production costs : Costs consumed during the production process, including raw material consumption, labor costs, etc.; Inventory holding costs : Costs of inventory storage, safekeeping and capital occupation; Responsiveness to market demand : Measures the system's response speed to changes in market demand; Inventory Consumption Rate :The speed at which inventory is consumed reflects inventory liquidity.

[0083] Among them, the reward function can be expressed as: ; in: is the impact weight of inventory deviation on reward; deviation is the deviation between the current inventory and the target inventory.

[0084] In some embodiments, the reward function can also be dynamically adjusted according to market changes and supply chain conditions, adjusting the weight coefficient of the reward function to guide the agent to optimize different goals. For example, when inventory is overstocked, the system may increase the weight of the inventory consumption rate to force the agent to speed up inventory turnover.

[0085] The collaboration between the reinforcement learning module and other modules (such as the multi-objective optimization module, nonlinear dynamics module, etc.) is the key to the efficient operation of this system. The reinforcement learning module continuously optimizes production plans, procurement strategies, and inventory management strategies, providing optimized decision inputs for multi-objective optimization and nonlinear dynamics models.

[0086] In the multi-objective optimization module, the optimization decisions provided by the reinforcement learning module will be used as input to help optimize the balance between multiple objectives such as production costs, inventory holding costs, and market demand response speed.

[0087] The nonlinear dynamics module helps simulate the dynamic changes of production and supply chain based on reinforcement learning optimized production scheduling, inventory management and procurement strategies, so as to predict future production, inventory, etc. and further optimize decision-making.

[0088] Multi-objective optimization module optimizes multiple objectives such as production cost, inventory holding cost, and capacity utilization to obtain the best decision-making solution; In this embodiment, the multi-objective optimization module accepts the optimization decision from the reinforcement learning module and the real-time data provided by the data acquisition and preprocessing module, and comprehensively optimizes the production plan, procurement strategy, inventory management and other decisions based on the multi-objective optimization algorithm. The optimized decision will directly affect the various indicators of production scheduling, inventory management and supply chain, such as production cost, inventory turnover rate and delivery cycle.

[0089] The core task of the multi-objective optimization module is to find a decision solution that achieves the best compromise between all objectives based on the relationship between different objectives in the system. This process involves integrating multiple optimization objectives into a unified objective function and solving the optimal solution through optimization algorithms.

[0090] In some embodiments, the multi-objective optimization module can use the following two common methods to perform optimization: Weighted Sum Method: The weighted sum method converts multiple objectives into a single objective problem by assigning weights to different optimization objectives, and then solves the problem. The weight of each objective reflects its relative importance to the overall optimization objective.

[0091] Specifically, assuming there is Optimization goals , the weighted sum method combines these The objective functions are combined into a single objective function , the formula is as follows: ; in: is the comprehensive objective function; For the An objective function (such as production cost, inventory cost, etc.); is the objective function The weight of represents the relative importance of the objective to the comprehensive optimization objective; To optimize the variable vector, it usually includes decision variables such as production plan, inventory level, purchase quantity, etc.

[0092] In this method, the weight coefficient It can be adjusted dynamically according to the importance of different goals. For example, when inventory management becomes the main goal, the weight of inventory cost can be increased and the weight of production cost can be reduced.

[0093] Pareto Optimization: As an option, if the relationship between the objectives is complex and cannot be simply weighted and processed, the Pareto Optimization method can be used for multi-objective optimization. The Pareto Optimization method refers to finding a solution among multiple objectives that cannot improve one objective without deteriorating other objectives.

[0094] Specifically, in Pareto optimization, the objective function is not transformed into a single objective problem through the weighted sum method, but rather a set of solutions is found such that for any solution , there is no other solution , which is better than , and outperforms If there is no such solution, then It is regarded as the Pareto optimal solution.

[0095] In multi-objective optimization, each objective function is associated with one or more constraints. Typically, the objective functions for production, inventory, and procurement are subject to resource limitations, production capacity, budget, and other constraints. Here are some common objective functions and their constraints: Production cost function: production objective function Aims to minimize production costs. Assume For the The unit cost of a product, is the production volume of the product, the production cost function can be expressed as: ; in: For the The unit production cost of a product; For the Production volume of a product.

[0096] Inventory consumption rate objective function: Inventory consumption rate Used to reflect the inventory turnover efficiency. By monitoring the inventory consumption rate, the system can adjust production and procurement plans to avoid excessive inventory backlogs: ; in: The amount of inventory consumed in the past period of time; The total inventory during that period.

[0097] Constraints: Multi-objective optimization is usually subject to some constraints. Common constraints include production capacity constraints, inventory capacity constraints, market demand, etc. The following are some common examples of constraints: Production capacity constraint: The production plan cannot exceed the production capacity: ; in, It is Production volume of the product; is the number of product types produced; It is the maximum capacity of a production facility, indicating the maximum quantity that the production facility can produce within a certain period of time.

[0098] Inventory capacity constraints; inventory cannot exceed the maximum capacity of the warehouse: ; in, It is Inventory of the product; is the total number of product categories; It is the maximum inventory capacity of a system (such as a warehouse).

[0099] The solution of multi-objective optimization usually uses heuristic algorithms (such as genetic algorithms, particle swarm optimization, etc.) or mathematical programming methods (such as linear programming, integer programming). For example, in production scheduling and procurement decision-making problems, genetic algorithms and particle swarm optimization methods search for optimal solutions by simulating the evolutionary process or particle motion process in nature.

[0100] Specifically, when using genetic algorithms, the optimization process continuously generates new solutions through operations such as selection, crossover, and mutation, and evaluates the fitness of the solutions, so as to finally obtain the optimal solution. In particle swarm optimization, the optimization process gradually searches for the optimal solution by simulating the movement of each particle in the solution space.

[0101] The collaboration between the multi-objective optimization module and the reinforcement learning module and the nonlinear dynamics module is crucial. The production, procurement, and inventory decisions optimized by the reinforcement learning module will be passed as input to the multi-objective optimization module. The optimization is performed through the weighted sum method or the Pareto optimal solution method, and finally the optimal solution that meets all the objective constraints is obtained.

[0102] In the reinforcement learning module, the agent makes optimization decisions based on real-time data, and these decisions are passed as input to the multi-objective optimization module; In the nonlinear dynamics module, the optimized production plan and inventory strategy will be used as input, and the nonlinear dynamics model will be used to further optimize the production scheduling and inventory management strategy to ensure the coordinated operation of all links.

[0103] Nonlinear dynamics module, which models the dynamic relationship between production, procurement and inventory based on nonlinear dynamics model to ensure coordinated operation of each link; In this embodiment, the nonlinear dynamics module uses the Lotka-Volterra model to describe the interdependence between production, inventory, and procurement. Based on real-time data and optimized decisions, the model can predict the future state of production, inventory, and procurement, and then adjust the decision. The following is the detailed construction of the model and the definition of the formula parameters.

[0104] The nonlinear dynamics module simulates the interactions between production, inventory, procurement and other links by using nonlinear system equations such as the Lotka-Volterra model. Traditional linear models fail to accurately reflect the complex feedback relationships between these links, so nonlinear dynamic models can more realistically reflect system behavior and feedback mechanisms.

[0105] In this embodiment, the Lotka-Volterra model is used to describe the dynamic relationship between production volume, inventory volume, and purchase volume. Assumptions: is the production volume; is the inventory; is the purchase volume; For market demand.

[0106] The Lotka-Volterra model uses the following equation to describe the dynamic changes between production, inventory, and procurement: Dynamic equation of production volume: ; : Production volume, expressed in time Production volume at the moment. : Production growth rate, which indicates the growth rate of production in unit time. : The interaction coefficient between production and inventory indicates the degree of influence of inventory on production. A negative value indicates an adverse impact on production when there is excess inventory. : Inventory quantity, expressed at time The inventory quantity at the moment. : The impact coefficient of market demand on production, indicating the positive impact of changes in market demand on production volume.

[0107] Inventory dynamic equation: ; in, : Inventory quantity, expressed at time The inventory quantity at the moment. : Inventory growth rate, which indicates the growth rate of inventory per unit time, which may be due to purchasing or production backlogs. : The mutual influence coefficient between inventory and production represents the feedback effect of inventory on production. A negative value represents the impact of excessive inventory on production.

[0108] Dynamic equation of purchase volume: ; in, : Purchase quantity, expressed in time Purchase volume at the time. : Purchase growth rate, which indicates the growth rate of purchase volume over time. : The interaction coefficient between procurement and inventory, which indicates the impact of inventory consumption on procurement. A negative value indicates the impact on procurement when there is inventory backlog.

[0109] As an option, the model can be further extended to take into account production capacity constraints or other factors. For example, assuming that the production capacity As a limiting factor, the production dynamic equation can be modified: Expanded production dynamic equation: ; in, : The maximum value of production capacity, indicating the upper limit of production capacity. : The adjustment factor of production capacity limit, which indicates the adjustment intensity when the production volume exceeds the maximum production capacity.

[0110] The nonlinear dynamics module not only makes predictions based on the current state, but also dynamically adjusts decisions through a real-time feedback mechanism. Each adjustment to a production, procurement, or inventory decision will affect the state of production, inventory, and procurement in the next time step. This dynamic feedback mechanism helps the system cope with real-time changes and ensures the optimal allocation of resources.

[0111] Specifically, during the production process, if the system detects a delay or insufficient inventory in a certain production link, the nonlinear dynamics module will make adjustments in advance by predicting future inventory and production needs. Based on historical data and current status predictions, the system will adjust procurement volume or production scheduling in a timely manner to respond to changes in market demand and internal bottlenecks.

[0112] The nonlinear dynamics module works closely with the aforementioned reinforcement learning module and multi-objective optimization module to ensure that the system's decision results are fully optimized. The decisions provided by the reinforcement learning module will be further verified and adjusted in the nonlinear dynamics model to ensure that the optimization decision not only conforms to the real-time status, but also adapts to future changes.

[0113] Specifically, in the reinforcement learning module, the agent optimizes decisions through historical data and feedback, and these decisions are passed as input to the nonlinear dynamics module. The nonlinear dynamics module uses dynamic models to predict future states and further adjust production plans, procurement plans, and inventory management strategies. Through this feedback loop, the system is able to maintain optimal resource allocation and decision-making in a constantly changing production and market environment.

[0114] The following are the parameter definitions for each formula in the nonlinear dynamics module to ensure that the meaning of each parameter is clear and consistent: Dynamic equation of production volume: ; in, : Production volume (unit: quantity); : Production growth rate (unit: 1 / time); : interaction coefficient between production and inventory (unit: 1 / (quantity × inventory)); : Inventory quantity (unit: quantity); : The influence coefficient of market demand on production (unit: 1 / (time × market demand)); :Market demand (unit: quantity).

[0115] Inventory dynamic equation: ; in, : Inventory quantity (unit: quantity); : Inventory growth rate (unit: 1 / time); : The mutual influence coefficient between inventory and production (unit: 1 / (inventory × production)); : Production volume (unit: quantity).

[0116] Dynamic equation of purchase volume: ; in, : Purchase quantity (unit: quantity); : Purchase growth rate (unit: 1 / time); : The interaction coefficient between purchase and inventory (unit: 1 / (purchase quantity × inventory quantity)); : Inventory quantity (unit: quantity).

[0117] Expanded production dynamic equation (considering production capacity constraints): ; in, : Production volume (unit: quantity); : Maximum production capacity (unit: quantity); : Adjustment factor for production capacity limit (unit: 1 / (quantity × maximum production capacity)).

[0118] The decision support module generates production plans, procurement plans, and inventory management strategies based on the output results of the reinforcement learning module and the multi-objective optimization module, and makes adjustments through real-time feedback; In this embodiment, the decision support and reporting module receives optimization decisions and real-time data from the reinforcement learning module, the multi-objective optimization module, the nonlinear dynamics module, etc., and organizes this information into easy-to-interpret reports for decision makers to refer to. These reports include not only the current status of the system, but also predictions and optimization suggestions for future trends. Through the reports, managers can quickly identify potential problems and take corresponding actions, thereby improving decision-making efficiency and reducing operational risks.

[0119] The core work of the decision support and reporting module is to integrate and display the optimization decisions, forecast results and system status data from each module. By comprehensively analyzing the data of production, inventory, procurement, sales and other links, the module generates detailed reports to display the operation status of each link and provide operational suggestions for managers. The system supports dynamic report generation, allowing decision makers to respond quickly based on real-time data.

[0120] During data transmission and processing, the module will work using the following steps: Data aggregation and processing: Obtain production, inventory, procurement and other related data from the reinforcement learning module, nonlinear dynamics module, etc., and aggregate and process these data.

[0121] Report generation: According to the set report format, various key indicators (such as production progress, inventory consumption, procurement needs, etc.) are visualized and various types of reports are generated.

[0122] Data analysis and trend forecasting: Combine historical data with real-time data to analyze future trends, generate forecast reports, and provide forward-looking support for decision-making.

[0123] Optimization decision suggestions: Based on the data displayed in the report and the decisions provided by the multi-objective optimization module, adjustment strategy suggestions are automatically generated.

[0124] The report module can customize the report content according to different management needs. Specifically, managers can choose to generate a detailed report for a specific link (such as production, inventory, procurement, etc.), or generate a comprehensive report covering multiple links. The report content includes but is not limited to: Production progress report: Displays the comparison between production plan and actual progress, helping managers evaluate production efficiency.

[0125] Inventory management report: displays information such as inventory quantity, consumption rate, inventory turnover, etc. to help managers adjust inventory strategies.

[0126] Procurement demand forecast report: Based on market demand and production plan, predict future procurement demand and adjust procurement strategy in advance.

[0127] In some embodiments, the report generation process may include the following steps: Input data integration: collect real-time data from each link, including status data from the production line, inventory data from the warehouse, demand data from the market, etc.

[0128] Data filtering and processing: remove irrelevant data, fill in missing data, and ensure the consistency and accuracy of input data.

[0129] Customized report generation: Generate relevant customized reports based on management needs, showing detailed data on production, inventory or procurement aspects of concern.

[0130] In order to improve the operability of the report, in this embodiment, the report module provides a powerful visualization function. Managers can intuitively view the current status and forecast of production, inventory and procurement through charts, trend charts, pie charts, etc. For example: Production efficiency chart: shows the comparison between planned production progress and actual production progress, helping managers identify production bottlenecks.

[0131] Inventory trend chart: displays real-time changes in inventory to help managers understand inventory dynamics.

[0132] Procurement demand forecast chart: Based on market demand and production conditions, a procurement demand forecast chart is generated to help managers formulate procurement strategies.

[0133] Specifically, during the presentation, data will be updated in real time in the form of dynamic charts, allowing managers to obtain the latest operational information. The dynamic update of the report can help managers flexibly adjust production, inventory and procurement strategies to optimize resource allocation.

[0134] Generally, the system can generate real-time reports and historical data comparison reports to provide managers with comprehensive information. These reports not only help managers understand the current system status, but also analyze trends and discover potential problems through historical data comparison, so as to make decisions and adjustments in advance.

[0135] Real-time reports: Display the current status of production, inventory and procurement, such as production volume, inventory consumption rate, order delivery time, etc., to help managers quickly understand the current business status.

[0136] Historical data comparison report: By comparing with historical data, it helps managers identify production bottlenecks, inventory backlogs and other problems, and make targeted adjustments. Through trend analysis, the system can help identify whether there are long-term production delays or inventory backlogs.

[0137] Specifically, the system can produce the following report types by comparing historical data with real-time data: Inventory change trend report: predict future inventory change trends based on historical inventory levels and current inventory data.

[0138] Production efficiency comparison report: Evaluate changes in production efficiency and production capacity by comparing the execution of production plans in different time periods.

[0139] Procurement strategy adjustment report: Based on historical procurement volume and actual consumption, evaluate the execution effect of the procurement plan and provide adjustment suggestions.

[0140] The decision support module not only generates data reports, but also generates strategy adjustment suggestions based on the output of the optimization algorithm. As an option, if the system predicts that there is a risk in a certain link (such as inventory shortages, production delays, etc.), the reporting module will automatically generate adjustment suggestions. These suggestions can cover the following: Suggestions for production plan adjustments: If the system predicts insufficient inventory, the reporting module can suggest increasing production volume or adjusting production scheduling.

[0141] Inventory replenishment suggestions: Based on inventory consumption rate and market demand forecast, the system can suggest advance purchases or adjust purchase quantities to ensure sufficient inventory.

[0142] Procurement adjustment suggestions: If the delivery time of certain suppliers is delayed, the reporting module can suggest changing suppliers or adjusting the procurement time.

[0143] The operation of the decision support and reporting module depends on the collaboration with other modules to ensure that the optimization decisions can be quickly fed back into the actual operation. The optimization decisions from the reinforcement learning module, nonlinear dynamics module and multi-objective optimization module will be passed as input to the reporting module. The reporting module generates relevant reports based on these decisions and provides adjustment suggestions.

[0144] In the reinforcement learning module, the agent makes decisions through learning and feedback, and the generated decisions such as production plans, inventory adjustments, and procurement plans will serve as inputs to the reporting module.

[0145] In the nonlinear dynamics module, based on the predicted production and inventory data, the reporting module will automatically generate strategic reports on production scheduling, inventory adjustment, etc.

[0146] In the multi-objective optimization module, by optimizing multiple objectives, the optimal decision provided by the system will be fed back to managers through the reporting module to generate targeted adjustment suggestions.

[0147] The following are commonly used formulas and parameter definitions in the report module to ensure that the meaning of each parameter is clear and consistent: Production progress calculation formula: ; in: is the production efficiency (unit: %), which indicates the ratio of the actual production progress to the planned progress; is the actual production volume (unit: quantity); is the planned production volume (unit: quantity).

[0148] Inventory consumption rate formula: ; in: is the inventory consumption rate (unit: %), indicating the rate of inventory consumption; It is the inventory consumed in a certain period of time (unit: quantity); It is the total inventory during this period (unit: quantity).

[0149] Procurement demand forecast formula: ; in: is the predicted purchase demand (unit: quantity); is the market demand (unit: quantity); is the inventory consumption rate (unit: %).

[0150] Please see attached Figure 2 , and also provides a ring network equipment production, supply and marketing integrated data processing method, comprising the following steps: S1. Collect relevant data on ring network equipment production, procurement, inventory, sales, etc. First of all, each link of ring network equipment, including production, procurement, inventory and sales, will generate a large amount of data. The system collects key data such as production progress, inventory quantity, procurement demand and sales orders in real time through integrated sensors, enterprise resource planning system (ERP), supply chain management system (SCM) and other information platforms. These data will become the basic input of subsequent optimization modules and provide an accurate basis for decision-making.

[0151] S2. Clean and standardize the collected data to ensure data consistency and quality. The collected data may be affected by noise, missing values ​​or inconsistencies during transmission, so it needs to be cleaned and standardized. Data cleaning includes removing duplicate data, filling missing values ​​and correcting abnormal data, while data standardization converts data in different formats and units into a unified format to ensure the effectiveness and accuracy of subsequent analysis. After this process, the consistency and quality of the data are effectively guaranteed.

[0152] S3. Dynamically optimize production plans, procurement plans, and inventory management based on reinforcement learning algorithms. The agent adjusts its decision-making strategy based on historical data and real-time feedback. After data preprocessing, the system uses reinforcement learning algorithms to dynamically optimize production plans, procurement plans, and inventory management strategies. The reinforcement learning agent evaluates the effects of different decisions through historical data and real-time feedback, and continuously adjusts its strategy within each decision cycle. Through trial and error learning, the agent gradually finds the optimal production scheduling, inventory adjustment, and procurement decisions, thereby improving the overall system's operating efficiency and resource utilization.

[0153] S4. Use a multi-objective optimization algorithm to optimize multiple objectives such as production cost, inventory cost and capacity utilization to obtain the best decision-making solution. The target weights of the multi-objective optimization are defined by the user. The system uses a multi-objective optimization algorithm to optimize multiple key objectives (such as production cost, inventory cost, capacity utilization, etc.). By assigning appropriate weights to each objective, the system can find the best compromise between these objectives. Users can adjust the weights of the objectives according to actual needs to achieve specific optimization goals. The optimization results provide the best decision-making solution for production, procurement and inventory management, ensuring the rational allocation and utilization of resources.

[0154] S5. Use nonlinear dynamic models to model the dynamic relationship between production, procurement and inventory, and adjust decisions based on feedback. In the nonlinear dynamic model, the system takes into account the mutual influence and feedback effects of each link by modeling the dynamic relationship between production, procurement and inventory. By predicting the dynamic changes of the system, the nonlinear dynamic model can identify potential problems and make adjustments. Based on these feedbacks, the system will further optimize production plans, inventory levels and procurement strategies to cope with changes in the market and production environment, and ensure the coordination and efficient operation of each link.

[0155] S6. Generate production plans, procurement plans and inventory management strategies based on the optimization results, and make decision adjustments through the real-time data update mechanism; Based on the above optimization process, the system will generate specific production plans, procurement plans and inventory management strategies. These decisions will be adjusted under the real-time data update mechanism to cope with changes in the external environment. For example, the system will automatically adjust production and procurement strategies according to real-time changes in market demand, production progress and inventory levels to ensure that the operation of each link is always in the optimal state.

[0156] S7. The system continuously adjusts the optimization plan through the feedback mechanism to ensure that all links are always in efficient coordination and optimal state. The system continuously adjusts the optimization plan through the feedback mechanism to ensure efficient coordination between all links. At the end of each cycle, the system will compare the actual execution results with the expected results and revise the optimization decisions. Through this continuous learning and adjustment, the system can adapt to the ever-changing market demand, production environment and supply chain conditions to ensure long-term efficient, flexible and sustainable operations.

[0157] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The integrated production, supply and marketing automation management system for ring network equipment is characterized by: include: Data collection module, used to collect data related to production, procurement, inventory and sales of ring network equipment in real time; Data preprocessing module, used to clean, standardize, fill missing values ​​and detect outliers on the collected data to ensure data quality; Reinforcement learning module, based on the collected real-time data and historical data, uses reinforcement learning algorithms to dynamically optimize production plans, procurement plans and inventory management; Multi-objective optimization module optimizes multiple objectives such as production cost, inventory holding cost, and capacity utilization to obtain the best decision-making solution; Nonlinear dynamics module, which models the dynamic relationship between production, procurement and inventory based on nonlinear dynamics model to ensure coordinated operation of each link; The decision support module generates production plans, procurement plans and inventory management strategies based on the output results of the reinforcement learning module and the multi-objective optimization module, and makes adjustments through real-time feedback.

2. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1 is characterized in that: The data acquisition module includes an Internet of Things device and an API interface, which is used to collect real-time data from the production line, warehouse and sales system of the ring network equipment. The data includes production progress, production capacity, raw material consumption, market demand, customer orders, and inventory consumption rate.

3. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1 is characterized in that: The reinforcement learning module trains the intelligent agent based on real-time data and historical data through Q learning or deep Q network algorithm, so that the intelligent agent can adaptively adjust the production plan, procurement plan and inventory management strategy according to environmental changes, and dynamically adjust the reward function in the reinforcement learning process through a real-time feedback mechanism to ensure real-time optimization of the intelligent agent strategy.

4. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1, characterized in that: The multi-objective optimization module uses a weighted sum method to weight multiple objectives and convert them into a single-objective optimization problem for solution. The objectives include production cost, inventory cost, and capacity utilization, where the weight of the objective function is customized by the user to achieve a trade-off between different objectives.

5. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1, characterized in that: The nonlinear dynamics module is dynamically adjusted through an automated management system, using the Lotka-Volterra model to simulate the interactions between production, procurement and inventory links, and to adjust the control strategies of each link based on real-time data. The model is used to dynamically optimize feedback decisions to ensure continuous coordination and efficient operation of each link.

6. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1, characterized in that: The decision support module includes a real-time data update module, which is used to adjust and optimize the production plan, procurement plan and inventory management strategy in real time according to the actual changes in production, procurement, sales and inventory, and continuously adjust the optimization plan according to the feedback mechanism to form a closed loop of continuous improvement and optimization.

7. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1, characterized in that: The data collection module is connected to a centralized data processing platform, which aggregates and classifies the collected production progress, inventory status, raw material consumption, and customer order data, and generates a real-time decision support report through preset report generation rules.

8. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1, characterized in that: The reinforcement learning module receives real-time data on market demand changes, supplier delivery dates and production capacity.

9. The integrated production, supply and marketing automated management system for ring network equipment according to claim 1, characterized in that: The multi-objective optimization module performs the following steps in sequence based on the input data of production cost, inventory holding cost, and capacity utilization rate collected in real time: Accepts user-defined objective weights and transforms multiple objectives into a single-objective optimization problem; Construct objective functions based on real-time and historical data and determine constraints; Obtain the global optimal solution through iterative optimization algorithm; Output optimized production plans, procurement plans and inventory management strategies, and adjust relevant parameters in real time to adapt to environmental changes.

10. A method for processing data of integrated production, supply and marketing for ring network equipment, applied to an integrated production, supply and marketing automation management system for ring network equipment as claimed in any one of claims 1 to 9, characterized in that: The following steps are involved: Collect data related to ring network equipment production, procurement, inventory and sales; Clean and standardize the collected data to ensure data consistency and quality; Dynamically optimize production plans, procurement plans, and inventory management based on reinforcement learning algorithms, and the agent adjusts its decision-making strategy based on historical data and real-time feedback; Use a multi-objective optimization algorithm to optimize multiple objectives such as production cost, inventory cost and capacity utilization to obtain the best decision-making solution, where the objective weights of the multi-objective optimization are defined by the user; Use nonlinear dynamic models to model the dynamic relationship between production, procurement and inventory, and adjust decisions based on feedback; Generate production plans, procurement plans and inventory management strategies based on optimization results, and make decision adjustments through real-time data update mechanism; The optimization plan is continuously adjusted through the feedback mechanism to ensure that all links are always in efficient coordination and optimal state.

Images