Supply chain allocation system and method based on big data

Abstract

The invention discloses a supply chain deployment system and method based on big data, and particularly relates to the technical field of supply chain deployment, and the system comprises a data collection module which is used for collecting the data of each link of a supply chain; a data storage and processing module; the demand prediction module is used for carrying out market demand prediction by using a time sequence analysis model and a regression analysis model; the intelligent allocation module is used for formulating an optimal supply chain allocation scheme by using an optimization algorithm; the logistics optimization module is used for carrying out optimization scheduling on logistics transportation routes and vehicles; the monitoring and feedback module is used for monitoring the operation state of each link of the supply chain in real time; according to the supply chain deployment system and method based on big data, the demand prediction accuracy is improved, through big data analysis and an advanced prediction algorithm, multiple influence factors are fully considered, the market demand can be predicted more accurately, the inventory overstock and stockout phenomena are reduced, and the inventory cost and opportunity cost of an enterprise are reduced.

Description

Technical Field

[0001] The present application relates to the field of supply chain allocation technology, and more specifically, to a supply chain allocation system and method based on big data. Background Art

[0002] In today's globalized and information-based business environment, supply chain management is crucial to the operation and development of enterprises. Traditional supply chain coordination methods, which rely primarily on manual experience and simple information systems, present numerous problems. For example, the inability to acquire and analyze large amounts of supply chain-related data in real time makes it difficult for enterprises to accurately predict market demand, resulting in frequent inventory backlogs and stockouts. In the logistics and distribution process, a lack of comprehensive information on transportation routes, vehicle status, and other aspects leads to high logistics costs and low distribution efficiency. Furthermore, data sharing and coordination between different links are poor, resulting in delays and errors in information transmission, which seriously affects the overall operational efficiency and responsiveness of the supply chain. With the continuous development of big data technology, how to effectively apply it to supply chain coordination to address these issues has become a topic that urgently needs research.

[0003] Therefore, to address the above problems, a supply chain allocation system and method based on big data are proposed. Summary of the Invention

[0004] The purpose of this application is to provide a supply chain deployment system and method based on big data.

[0005] This application provides a supply chain allocation system and method based on big data, which adopts the following technical solutions: A supply chain coordination system based on big data, the supply chain coordination system comprising: Data acquisition module: used to collect data from all links of the supply chain, and the output end of the data acquisition module is electrically connected to the input end of the data storage and processing module; Data storage and processing module: used to clean, convert and integrate the data collected by the data acquisition module, remove duplicate, erroneous and incomplete data, and store the processed data in a distributed database; Demand forecasting module: Based on the data provided by the data storage and processing module, it uses time series analysis models and regression analysis models to forecast market demand; Intelligent allocation module: Based on the forecast results of the demand forecast module, it uses optimization algorithms to formulate the optimal supply chain allocation plan; Logistics Optimization Module: Based on the supply chain allocation plan developed by the intelligent allocation module, it uses the path optimization algorithm and vehicle scheduling algorithm in combination with logistics data to optimize the logistics transportation routes and vehicles; Monitoring and feedback module: monitors the operating status of each link in the supply chain in real time, and compares and analyzes the monitoring data with preset indicators.

[0006] Furthermore, the data of each link in the data acquisition module include but are not limited to the supplier's raw material supply data, production data of the production link, sales data of the sales link, logistics data of the logistics link and market environment data, and the data acquisition module is designed to connect with the information system of each link in the supply chain through a data interface, and use Internet of Things devices to collect data in real time.

[0007] Furthermore, the data storage and processing module uses a hash algorithm to identify and delete duplicate data, corrects erroneous data through preset numerical range verification rules and logical relationship verification rules, and uses a mean filling algorithm or a regression prediction filling algorithm to complete incomplete data. In the conversion process, timestamps, units, and codes in different formats are uniformly standardized, and key information is extracted from unstructured data through natural language processing technology to convert it into structured data.

[0008] Furthermore, the output end of the data storage and processing module is electrically connected to the input end of the demand forecasting module, and the time series analysis model in the demand forecasting module implements the following strategy: Forget Gate: ; Input Gate: ; Cell status update: ; Cell status: ; Input Gate: ; Hidden state: ; in, is the sigmoid activation function, tanh is the hyperbolic tangent activation function, 、 、 、 is the weight matrix, 、 、 、 is the bias vector, Indicates that the previous moment is hidden With the current input Splicing, is the updated cell state, For the current hidden state, in the supply chain demand forecast, the historical sales data, inventory turnover rate, and holiday distribution time series data are encoded in sequence as input After multiple layers of iterative learning, the model outputs demand forecasts for the future. By comprehensively considering various relevant features such as historical sales data, seasonal factors, and promotional activities, the time series analysis model can fully learn the complex time-dependent patterns and underlying laws in the data. During the training process, network parameters are continuously adjusted to optimize model performance, ensuring that the forecast results more accurately reflect the actual changes in market demand, providing a reliable demand forecast foundation for subsequent supply chain allocation decisions. The regression analysis model uses the forest regression algorithm, specifically: random forest regression consists of multiple decision trees, and the final prediction value is obtained by averaging the prediction results of multiple decision trees; For a single decision tree, when a node is split, the optimal splitting feature and threshold are selected based on minimizing the mean square error (MSE). The mean square error formula is: ,in is the sample size, is the true value, To predict the value, in demand forecasting, historical sales data, seasonal factors, and promotion-related features are used as input to train multiple decision trees. Finally, the prediction results of all decision trees are averaged to obtain the final market demand forecast result.

[0009] Furthermore, the optimization algorithm in the intelligent allocation module adopts a combination of genetic algorithm and particle swarm optimization algorithm, and the genetic algorithm is optimized by simulating natural selection and genetic process. The specific steps are as follows: When determining the raw material procurement plan, binary coding or real number coding is used to encode the procurement quantity, procurement time, and procurement supplier decision variables into chromosomes, with minimizing procurement costs and meeting demand as the objective function. The objective function can be expressed as , where m is the number of purchase batches, is the unit price of the i-th batch of raw materials, is the purchase quantity of batch i, is the transportation cost of batch i, To select the additional cost of the jth supplier, a better chromosome is selected according to the fitness function through the selection operation; The crossover operation exchanges some genes of two chromosomes to produce new chromosomes; The mutation operation randomly changes certain genes of the chromosome to increase population diversity; After multiple generations of evolution, the optimal procurement solution is obtained.

[0010] Furthermore, each particle in the particle swarm optimization algorithm searches for the optimal solution in the solution space. The position of the particle represents a set of decision variables, and the speed determines the direction and distance of the particle movement. For example, the particle is based on its own historical optimal position. and the global optimal position Update speed and position. The update formula is as follows: Speed ​​update: + ; Location Updates: ; in, and are the d-th dimension components of the velocity and position of particle i at time t, is the inertia weight, and is the learning factor, and for A random number between is the historical optimal position of particle i dimensional component, The global optimal position Dimensional component; When formulating a production plan, production batches, production time, and production task allocation are used as particle position variables. With the goal of minimizing production costs and maximizing production efficiency, the optimal production plan is searched by continuously updating the particle speed and position.

[0011] Furthermore, the path optimization algorithm of the logistics optimization module adopts the Dijkstra algorithm to solve the single-source shortest path, specifically: Design , where V is the vertex set, E is the edit point, the source point is s, and each edge has a non-negative weight ; Maintain a distance array , used to record the shortest distance from the source point to each vertex, initially , the remaining vertices The value is set to infinity; In each iteration, select the unvisited vertices The vertex with the smallest value , then update its adjacent vertices of The update formula is: ; In logistics transportation route optimization, the transportation nodes are regarded as vertices, the road sections between the nodes are regarded as edges, and the transportation time or cost of the road sections are regarded as the weights of the edges. The shortest transportation routes from the starting point to each destination are calculated by the Dijkstra algorithm. The vehicle scheduling algorithm of the logistics optimization module adopts the vehicle routing problem algorithm with time window VRPTW. Under the conditions of meeting vehicle capacity constraints and customer time window constraints, the vehicle's driving route is determined to minimize the total transportation cost. Specifically: The cost function is: ,in, is an arc set, is the transportation cost from node i to node j, is the decision variable. If the vehicle moves from node i to node j, ,otherwise ; and are the early penalty cost and late penalty cost respectively, and are the early arrival and late arrival times of the vehicle at node i, respectively; By constructing an initial solution and then optimizing it using a simulated annealing algorithm or a tabu search algorithm, the optimal vehicle scheduling solution is obtained.

[0012] Furthermore, the operating status of each link in the monitoring and feedback module includes production progress, inventory level, logistics and transportation conditions, and supplier delivery conditions. When an abnormal situation occurs, an early warning message is issued in a timely manner, and the abnormal situation is fed back to the intelligent allocation module. The intelligent allocation module adjusts and optimizes the supply chain allocation plan based on the feedback information, forming a closed-loop supply chain management system.

[0013] Furthermore, a supply chain allocation method based on big data includes the following steps: Collect data from all links of the supply chain through the data collection module; Use the data storage and processing module to clean, convert, integrate and store the collected data, and store the processed data in a distributed database; Use machine learning algorithms and deep learning models in the demand forecasting module to forecast market demand based on processed data; With the help of the intelligent allocation module, based on the demand forecast results, combined with inventory data and supplier supply capacity data, time series analysis models and regression analysis models are used to predict market demand and formulate supply chain allocation plans; Utilize the logistics optimization module to optimize the logistics transportation routes and vehicles according to the material distribution plan and logistics data, and combine the logistics data with the path optimization algorithm and vehicle scheduling algorithm; monitor the supply chain operation status in real time through the monitoring and feedback module, feedback information in case of abnormalities, and the intelligent allocation module adjusts and optimizes the allocation plan accordingly.

[0014] The technical effects and advantages of this application are: Compared with existing technologies, this big data-based supply chain allocation system and method improves the accuracy of demand forecasting: through big data analysis and advanced forecasting algorithms, it fully considers multiple influencing factors, can more accurately predict market demand, reduce inventory backlogs and out-of-stock phenomena, and reduce the company's inventory costs and opportunity costs.

[0015] Achieve intelligent allocation: Utilize intelligent optimization algorithms to comprehensively consider various factors in the supply chain, formulate the best allocation plan, improve the utilization efficiency of supply chain resources, achieve reasonable allocation of resources, and enhance the competitiveness of enterprises.

[0016] Optimize logistics and distribution: Through optimized scheduling of logistics transportation routes and vehicles, real-time monitoring and dynamic adjustments, we can improve logistics and distribution efficiency, reduce logistics costs, and enhance customer satisfaction.

[0017] Enhance supply chain stability: Consider the uncertainties in the supply chain, develop emergency plans, and use monitoring and feedback mechanisms to promptly identify and resolve problems, ensure the stable operation of the supply chain, and reduce losses to enterprises caused by supply chain disruptions.

[0018] Improve the scientific nature of decision-making: Based on the comprehensive and accurate information provided by big data analysis, provide a scientific basis for corporate decision-making, help corporate managers make more informed decisions, and promote the sustainable development of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic diagram of the system flow of this application; Figure 2 Schematic diagram of the method flow of this application.

[0020] The accompanying drawings are marked as: 1. Data acquisition module; 2. Data storage and processing module; 3. Demand forecasting module; 4. Intelligent allocation module; 5. Logistics optimization module; 6. Monitoring and feedback module. DETAILED DESCRIPTION

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

[0022] Example 1 like Figures 1 to 2 A supply chain coordination system based on big data is shown, and the supply chain coordination system includes: Data acquisition module 1: used to collect data from various links of the supply chain, and the output end of the data acquisition module 1 is electrically connected to the input end of the data storage and processing module 2; The data of each link in the data collection module 1 include but are not limited to the supplier's raw material supply data (such as supply time, quality, price, etc.), the production data of the production link (such as production progress, equipment operation status, raw material consumption, etc.), the sales data of the sales link (such as sales orders, sales trends, customer demand preferences, etc.), the logistics data of the logistics link (such as transport vehicle location, transport route, transport time, cargo status, etc.) and market environment data (such as macroeconomic data, industry dynamics, competitor information, etc.), and the data collection module 1 is designed to connect with the information systems of each link in the supply chain (such as the supplier's ERP system, the enterprise's internal MES system, the sales platform's CRM system, the logistics company's logistics management system, etc.) through data interface docking, and use Internet of Things devices (such as sensors, RFID tags, etc.) to collect data in real time.

[0023] Data storage and processing module 2: used to clean, convert and integrate the data collected by the data acquisition module, remove duplicate, erroneous and incomplete data, and store the processed data in a distributed database; The data storage and processing module 2 uses a hash algorithm to identify and delete duplicate data, corrects erroneous data through preset numerical range verification rules and logical relationship verification rules, and uses a mean filling algorithm or a regression prediction filling algorithm to complete incomplete data. In the conversion process, timestamps, units, and codes in different formats are standardized and processed uniformly, and key information is extracted from unstructured data through natural language processing technology to convert it into structured data.

[0024] Demand forecasting module 3: Based on the data provided by data storage and processing module 2, the time series analysis model and regression analysis model are used to forecast market demand; The output terminal of the data storage and processing module 2 is electrically connected to the input terminal of the demand forecasting module 3, and the time series analysis model in the demand forecasting module 3 implements the following strategy: Forget Gate: ; Input Gate: ; Cell status update: ; Cell status: ; Input Gate: ; Hidden state: ; in, is the sigmoid activation function, tanh is the hyperbolic tangent activation function, 、 、 、 is the weight matrix, 、 、 、 is the bias vector, Indicates that the previous moment is hidden With the current input Splicing, is the updated cell state, For the current hidden state, in the supply chain demand forecast, the historical sales data, inventory turnover rate, and holiday distribution time series data are encoded in sequence as input After multiple layers of iterative learning, the model outputs demand forecasts for the future. By comprehensively considering various relevant features such as historical sales data, seasonal factors, and promotional activities, the time series analysis model can fully learn the complex time-dependent patterns and underlying laws in the data. During the training process, network parameters are continuously adjusted to optimize model performance, ensuring that the forecast results more accurately reflect the actual changes in market demand, providing a reliable demand forecast foundation for subsequent supply chain allocation decisions. The regression analysis model uses the forest regression algorithm, specifically: random forest regression consists of multiple decision trees, and the final prediction value is obtained by averaging the prediction results of multiple decision trees; For a single decision tree, when a node is split, the optimal splitting feature and threshold are selected based on minimizing the mean square error (MSE). The mean square error formula is: ,in is the sample size, is the true value, To predict the value, the module uses historical sales data, seasonal factors, and promotional features as input to train multiple decision trees. Finally, the prediction results of all decision trees are averaged to obtain the final market demand forecast. This module generates accurate demand forecasts through a comprehensive analysis of multiple factors affecting demand, providing a basis for decision-making in the intelligent allocation module.

[0025] Intelligent Allocation Module 4: Based on the forecast results of Demand Forecast Module 3, it uses optimization algorithms to formulate the optimal supply chain allocation plan; The optimization algorithm in the intelligent allocation module 4 adopts a combination of genetic algorithm and particle swarm optimization algorithm, and the genetic algorithm is optimized by simulating natural selection and genetic process. The specific steps are as follows: When determining the raw material procurement plan, binary coding or real number coding is used to encode the procurement quantity, procurement time, and procurement supplier decision variables into chromosomes, with minimizing procurement costs and meeting demand as the objective function. The objective function can be expressed as , where m is the number of purchase batches, is the unit price of the i-th batch of raw materials, is the purchase quantity of batch i, is the transportation cost of batch i, To select the additional cost of the jth supplier, a better chromosome is selected according to the fitness function through the selection operation; The crossover operation exchanges some genes of two chromosomes to produce new chromosomes; The mutation operation randomly changes certain genes of the chromosome to increase population diversity; After multiple generations of evolution, the optimal procurement solution is obtained.

[0026] In the particle swarm optimization algorithm, each particle searches for the optimal solution in the solution space. The position of the particle represents a set of decision variables, and the speed determines the direction and distance of the particle movement. For example, based on its own historical optimal position and the global optimal position Update speed and position. The update formula is as follows: Speed ​​update: + ; Location Updates: ; in, and are the d-th dimension components of the velocity and position of particle i at time t, is the inertia weight, and is the learning factor, and for A random number between is the historical optimal position of particle i dimensional component, The global optimal position Dimensional component; When formulating a production plan, production batches, production times, and production task allocations are used as particle position variables. With the goal of minimizing production costs and maximizing production efficiency, the module continuously updates particle speeds and positions to search for the optimal production plan. This module also considers uncertainties in the supply chain, such as supplier delivery delays, production failures, and market demand fluctuations, and develops corresponding contingency plans to ensure stable supply chain operations. Logistics Optimization Module 5: Based on the supply chain allocation plan developed by Intelligent Allocation Module 4, it uses path optimization algorithms and vehicle scheduling algorithms in combination with logistics data to optimize logistics transportation routes and vehicles. The path optimization algorithm of the logistics optimization module 5 uses the Dijkstra algorithm to solve the shortest path of a single source, specifically: Design , where V is the vertex set, E is the edit point, the source point is s, and each edge has a non-negative weight ; Maintain a distance array , used to record the shortest distance from the source point to each vertex, initially , the remaining vertices The value is set to infinity; In each iteration, select the unvisited vertices The vertex with the smallest value , then update its adjacent vertices of The update formula is: ; In logistics transportation route optimization, the transportation nodes are regarded as vertices, the road sections between the nodes are regarded as edges, and the transportation time or cost of the road sections are regarded as the weights of the edges. The shortest transportation routes from the starting point to each destination are calculated by the Dijkstra algorithm. The vehicle scheduling algorithm of the logistics optimization module 5 adopts the vehicle routing problem algorithm with time window VRPTW. Under the conditions of meeting vehicle capacity constraints and customer time window constraints, the vehicle's driving route is determined to minimize the total transportation cost. Specifically, The cost function is: ,in, is an arc set, is the transportation cost from node i to node j, is the decision variable. If the vehicle moves from node i to node j, ,otherwise ; and are the early penalty cost and late penalty cost respectively, and are the early arrival and late arrival times of the vehicle at node i, respectively; By constructing an initial solution and then optimizing it using a simulated annealing algorithm or a tabu search algorithm, the optimal vehicle dispatching plan is obtained. During the logistics transportation process, GPS positioning technology and IoT devices monitor the location, speed, and cargo status of transport vehicles in real time. Transport routes and dispatching plans are dynamically adjusted based on actual conditions to improve logistics distribution efficiency and reduce logistics costs. Monitoring and Feedback Module 6: Real-time monitoring of the operating status of each link in the supply chain, and comparative analysis of the monitoring data with preset indicators.

[0027] The operating status of each link in the monitoring and feedback module 6 includes production progress, inventory level, logistics and transportation conditions, and supplier delivery conditions. When an abnormal situation occurs, an early warning message is issued in a timely manner, and the abnormal situation is fed back to the intelligent allocation module 4. The intelligent allocation module 4 adjusts and optimizes the supply chain allocation plan based on the feedback information, forming a closed-loop supply chain management system.

[0028] Example 2 A supply chain allocation method based on big data includes the following steps: Collect data from all links of the supply chain through data collection module 1; The data storage and processing module 2 is used to clean, convert, integrate and store the collected data, and store the processed data in a distributed database; Use machine learning algorithms and deep learning models in the demand forecasting module to forecast market demand based on processed data; With the help of intelligent allocation module 4, based on the demand forecast results, combined with inventory data and supplier supply capacity data, time series analysis model and regression analysis model are used to predict market demand and formulate supply chain allocation plans; Utilize the logistics optimization module 5 to optimize the logistics transportation routes and vehicles according to the material distribution plan and logistics data, and combine the logistics data with the path optimization algorithm and vehicle scheduling algorithm; The monitoring and feedback module 6 monitors the supply chain operation status in real time, and feedback information is fed back when an abnormality occurs. The intelligent allocation module adjusts and optimizes the allocation plan accordingly.

[0029] Finally: The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A supply chain deployment system based on big data, characterized by: The supply chain deployment system includes: Data acquisition module (1): used to collect data from various links of the supply chain, and the output end of the data acquisition module (1) is electrically connected to the input end of the data storage and processing module (2); Data storage and processing module (2): used to clean, convert and integrate the data collected by the data acquisition module, remove duplicate, erroneous and incomplete data, and store the processed data in a distributed database; Demand forecasting module (3): Based on the data provided by the data storage and processing module (2), the market demand forecast is performed using the time series analysis model and the regression analysis model; Intelligent allocation module (4): Based on the forecast results of the demand forecast module (3), it uses the optimization algorithm to formulate the optimal supply chain allocation plan; Logistics Optimization Module (5): Based on the supply chain allocation plan developed by the Intelligent Allocation Module (4), the logistics data is combined with the path optimization algorithm and vehicle scheduling algorithm to optimize the logistics transportation routes and vehicles; Monitoring and feedback module (6): monitors the operating status of each link in the supply chain in real time, and compares and analyzes the monitoring data with the preset indicators.

2. The big data-based supply chain coordination system according to claim 1, characterized in that: The data of each link in the data acquisition module (1) include but are not limited to the supplier's raw material supply data, production data of the production link, sales data of the sales link, logistics data of the logistics link and market environment data, and the data acquisition module (1) is configured to connect with the information system of each link in the supply chain through a data interface and use Internet of Things devices to collect data in real time.

3. The big data-based supply chain coordination system according to claim 1, characterized in that: The data storage and processing module (2) uses a hash algorithm to identify and delete duplicate data, corrects erroneous data through preset value range verification rules and logical relationship verification rules, and uses a mean filling algorithm or a regression prediction filling algorithm to complete incomplete data. In the conversion process, timestamps, units, and codes in different formats are standardized and converted into structured data by extracting key information from unstructured data through natural language processing technology.

4. The big data-based supply chain coordination system according to claim 3, characterized in that: The output end of the data storage and processing module (2) is electrically connected to the input end of the demand forecasting module (3), and the time series analysis model in the demand forecasting module (3) executes the following strategy: Forget Gate: ; Input Gate: ; Cell status update: ; Cell status: ; Input Gate: ; Hidden state: ; in, is the sigmoid activation function, tanh is the hyperbolic tangent activation function, 、 、 、 is the weight matrix, 、 、 、 is the bias vector, Indicates hiding the previous state With the current input Splicing, is the updated cell state, For the current hidden state, in the supply chain demand forecast, the historical sales data, inventory turnover rate, and holiday distribution time series data are encoded in sequence as input , after multiple layers of iterative learning, outputs the demand forecast value at the future moment; The regression analysis model uses the forest regression algorithm, specifically: random forest regression consists of multiple decision trees, and the final prediction value is obtained by averaging the prediction results of multiple decision trees; For a single decision tree, when a node is split, the optimal splitting feature and threshold are selected based on minimizing the mean square error (MSE). The mean square error formula is: ,in is the sample size, is the true value, To predict the value, in demand forecasting, historical sales data, seasonal factors, and promotion-related features are used as input to train multiple decision trees. Finally, the prediction results of all decision trees are averaged to obtain the final market demand forecast result.

5. The big data-based supply chain coordination system according to claim 1, characterized in that: The optimization algorithm in the intelligent allocation module (4) adopts a combination of genetic algorithm and particle swarm optimization algorithm, and the genetic algorithm is optimized by simulating natural selection and genetic process. The specific steps are: When determining the raw material procurement plan, binary coding or real number coding is used to encode the procurement quantity, procurement time, and procurement supplier decision variables into chromosomes, with minimizing procurement costs and meeting demand as the objective function. The objective function can be expressed as , where m is the number of purchase batches, is the unit price of the i-th batch of raw materials, is the purchase quantity of batch i, is the transportation cost of batch i, To select the additional cost of the jth supplier, a better chromosome is selected according to the fitness function through the selection operation; The crossover operation exchanges some genes of two chromosomes to produce new chromosomes; The mutation operation randomly changes certain genes of the chromosome to increase population diversity; After multiple generations of evolution, the optimal procurement solution is obtained.

6. The big data-based supply chain coordination system according to claim 1, characterized in that: In the particle swarm optimization algorithm, each particle searches for the optimal solution in the solution space. The position of the particle represents a set of decision variables, and the speed determines the direction and distance of the particle movement. For example, the particle searches for the optimal solution based on its own historical optimal position. and the global optimal position Update speed and position. The update formula is as follows: Speed ​​update: + ； Location Updates: ; in, and are the d-th dimension components of the velocity and position of particle i at time t, is the inertia weight, and is the learning factor, and for A random number between is the historical optimal position of particle i dimensional component, The first position of the global optimal Dimensional component; When formulating a production plan, production batches, production time, and production task allocation are used as particle position variables. With the goal of minimizing production costs and maximizing production efficiency, the optimal production plan is searched by continuously updating the particle speed and position.

7. The big data-based supply chain coordination system according to claim 7, characterized in that: The path optimization algorithm of the logistics optimization module (5) adopts the Dijkstra algorithm to solve the single-source shortest path, specifically: Design , where V is the vertex set, E is the edit point, the source point is s, and each edge has a non-negative weight ; Maintain a distance array , used to record the shortest distance from the source point to each vertex, initially , the remaining vertices The value is set to infinity; In each iteration, select the unvisited vertices The vertex with the smallest value , then update its adjacent vertices of The update formula is: ; In logistics transportation route optimization, the transportation nodes are regarded as vertices, the road sections between the nodes are regarded as edges, and the transportation time or cost of the road sections are regarded as the weights of the edges. The shortest transportation routes from the starting point to each destination are calculated by the Dijkstra algorithm. The vehicle scheduling algorithm of the logistics optimization module (5) adopts the vehicle routing problem algorithm with time window VRPTW, which determines the vehicle's driving route to minimize the total transportation cost under the conditions of meeting vehicle capacity constraints, customer time window constraints, etc., specifically: The cost function is: ,in, is an arc set, is the transportation cost from node i to node j, is the decision variable. If the vehicle moves from node i to node j, ,otherwise ; and are the early penalty cost and late penalty cost respectively, and are the early arrival and late arrival times of the vehicle at node i, respectively; By constructing an initial solution and then optimizing it using a simulated annealing algorithm or a tabu search algorithm, the optimal vehicle scheduling solution is obtained.

8. The big data-based supply chain coordination system according to claim 1, characterized in that: The operating status of each link in the monitoring and feedback module (6) includes production progress, inventory level, logistics and transportation status, and supplier delivery status. When an abnormal situation occurs, an early warning message is issued in a timely manner, and the abnormal situation is fed back to the intelligent allocation module (4). The intelligent allocation module (4) adjusts and optimizes the supply chain allocation plan based on the feedback information, forming a closed-loop supply chain management system.

9. A supply chain deployment method based on big data, characterized by: The big data-based supply chain coordination system according to any one of claims 1 to 9 is applied, comprising the following steps: Collect data from all links of the supply chain through the data collection module (1); Using the data storage and processing module (2) to clean, transform, integrate and store the collected data, and store the processed data in a distributed database; Use the machine learning algorithms and deep learning models in the demand forecasting module (3) to forecast market demand based on the processed data; With the help of the intelligent allocation module (4), based on the demand forecast results, combined with inventory data and supplier supply capacity data, the time series analysis model and regression analysis model are used to predict market demand and formulate a supply chain allocation plan; Utilize the logistics optimization module (5) to optimize the logistics transportation routes and vehicles according to the material distribution plan and logistics data, and combine the logistics data with the path optimization algorithm and vehicle scheduling algorithm; The monitoring and feedback module (6) monitors the supply chain operation status in real time, and provides feedback when anomalies occur. The intelligent allocation module then adjusts and optimizes the allocation plan accordingly.

Images