An order distribution system and method for multi-warehouse supply of production raw materials

Abstract

The present invention relates to the technical field of collaborative operation of production systems and logistics systems in intelligent manufacturing scenarios, and discloses an order distribution system for multi-warehouse supply of production raw materials. The data analysis module of the system extracts data from data sources such as production plans and warehouse information to generate production raw material demand orders; the solution solving module calculates the optimal outbound warehouse and quantity based on a multi-objective optimization model, taking into account factors such as inventory, delivery time, and equipment load rate; the allocation execution module dispatches logistics equipment to complete the outbound task based on the optimization results; the monitoring and prediction module uses a long-short-term memory network to learn and predict historical and real-time order data, and dynamically adjusts the allocation strategy; the user interaction module provides a human-computer interaction interface. This system realizes efficient and economical JIT supply of production raw materials through intelligent means, significantly improves logistics efficiency and resource utilization, and provides reliable technical support for production and logistics collaboration in intelligent manufacturing scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of collaborative operation of production systems and logistics systems in intelligent manufacturing scenarios, and specifically to an order distribution system and method for multi-warehouse supply of production raw materials. Background Art

[0002] In smart manufacturing scenarios, the efficient supply of production raw materials is a key link in ensuring the stable operation of the production system. Normally, automated logistics systems deliver production raw materials to the vicinity of production machines in a JIT (Just-In-Time) manner to meet the real-time needs of the production line. However, in actual applications, due to the capacity limitations of a single automated warehouse, regional distribution differences, and the diverse needs of production equipment, materials of the same specification may need to be stored in multiple warehouses. For example, some materials may be stored in both a stacker crane pallet warehouse and a multi-material box warehouse. Although this multi-warehouse storage model solves the problem of insufficient capacity in a single warehouse, it also brings new technical challenges: how to select a suitable warehouse to process orders in a multi-warehouse environment to achieve efficient material outbound and distribution.

[0003] Currently, traditional order allocation methods rely primarily on manual assignment or pre-setting of simple outbound delivery rules, such as selecting the warehouse with the largest inventory or the warehouse closest to the warehouse. While these methods are simple to implement, they present the following technical issues:

[0004] Insufficient single-factor consideration: Existing methods typically focus solely on a single factor (such as inventory level or regional distribution), while ignoring the combined impact of other important factors, such as material delivery time, transportation costs, and equipment load factor. This single-dimensional decision-making approach makes it difficult to achieve optimal logistics efficiency and cost control.

[0005] Unbalanced resource utilization: Due to a lack of overall optimized scheduling of multi-warehouse resources, existing allocation strategies often lead to overloading of some warehouses while leaving others idle. This not only reduces the utilization of equipment resources but can also cause logistics bottlenecks, impacting overall production efficiency.

[0006] Lack of dynamic adaptability: Existing methods typically use fixed allocation rules that are unable to dynamically adjust to real-time changes in production order demand, warehouse inventory status, and equipment operation. This static allocation approach is inflexible in complex production environments and struggles to adapt to fluctuations in order demand and changes in warehouse status.

[0007] Lack of forecasting capabilities: Traditional methods lack the ability to predict future order demand and are unable to optimize order allocation strategies in advance, which may lead to excessive logistics pressure during peak periods or waste of resources during low periods.

[0008] In summary, current technical solutions have significant limitations in multi-warehouse environments and are unable to meet the demand for efficient and economical production raw material supply in intelligent manufacturing scenarios. Therefore, an order allocation system and method that can comprehensively consider multiple factors, dynamically adjust allocation strategies, and possess predictive capabilities is urgently needed to address these technical issues and improve the collaborative operation efficiency of production and logistics systems. Summary of the Invention

[0009] In order to solve the above problems, the present invention provides a multi-warehouse order distribution system that can comprehensively consider multiple factors such as inventory level, material delivery time, load rate of handling equipment in real time, and monitor the execution of order distribution in real time. Based on real-time and historical order data, it predicts future order demand, manually adjusts the multi-objective optimization model parameters, and optimizes the order distribution results to achieve efficient and economical JIT supply of production raw materials in intelligent manufacturing scenarios.

[0010] The technical solution adopted in the present invention is:

[0011] An order distribution system for multi-warehouse supply of production raw materials, comprising a data analysis module, a solution solving module, a distribution execution module, and a monitoring and prediction module connected in sequential data transmission, and a user interaction module connected in data transmission with the data analysis module, the solution solving module, the distribution execution module, and the monitoring and prediction module;

[0012] The data analysis module is responsible for data extraction, cleaning, standardization and calculation, and generates production raw material demand orders; the solution solving module solves the optimal outbound warehouse and quantity based on a multi-objective optimization model; the allocation execution module allocates the same raw materials in the order to different warehouses based on the results of the solution solving module, and dispatches logistics equipment to complete the outbound operation; the monitoring and prediction module monitors the order execution status and warehouse status in real time, predicts future order demand based on historical and real-time data, and optimizes the allocation strategy; the user interaction module provides a human-computer interaction interface, supports manual intervention and parameter adjustment, and optimizes order allocation results.

[0013] Furthermore, the data analysis module extracts data from a predetermined data source according to a set time point or is manually triggered to generate a production raw material data set; the extracted data is then anomaly identified and re-acquired when abnormal data is found;

[0014] For normal data, standardization conversion is performed to unify the data format and unit; then the standardized data is structured and organized according to the preset logical relationship to form a clear data structure; finally, based on the set correspondence between data, the structured data is calculated to generate production raw material demand orders.

[0015] Furthermore, the production raw material data set includes: production plan, product production time, production quantity, product BOM table, warehouse information, and material specification information, material quantity information and material packaging information of the corresponding warehouse.

[0016] Furthermore, in the production raw material multi-warehouse supply scheduling and distribution system, the production raw material demand order is one of the core data; each order corresponds to a raw material, which may be shipped from one or more warehouses and eventually supplied to a production unit for use.

[0017] Furthermore, the solution solving module defines four constraints with the goal of minimizing the comprehensive cost of production raw material supply: a first constraint, a second constraint, a third constraint, and a fourth constraint. The solution solving module uses the production raw material demand order data generated by the data analysis module as input and determines the shipping warehouse and corresponding quantity of each production raw material by solving a multi-objective optimization model.

[0018] The calculation formula of the multi-objective optimization model is:

[0019]

[0020] Where, It represents the comprehensive cost of supplying raw materials from multiple warehouses; Indicates the total number of orders, that is, the number of production raw material demand orders that need to be allocated; Indicates the total number of warehouses, that is, the number of warehouses available for selection; Indicates the index of the order, , used to identify the orders; Represents the index of the warehouse, , used to identify the warehouses; Represents a decision variable, which takes a value of 0 or 1. Indicates order By warehouse deal with, Indicates order Not by warehouse deal with; Represents the weight coefficients of different cost factors, satisfying , is the weight of inventory-related costs, is the weight of the material transportation time-related cost, The weight of the costs associated with the loading rate of the handling equipment; Indicates order Materials in the warehouse Inventory in Indicates order Materials in the warehouse The time required to ship from the warehouse to the destination; Represents a warehouse Loading rate of handling equipment; It means to perform a double summation on all orders and warehouses, that is, to comprehensively consider the distribution relationship between each order and each warehouse.

[0021] Furthermore, the first constraint is: each order can only be processed by one warehouse;

[0022] The calculation formula is:

[0023]

[0024] The second constraint is: the warehouse inventory must meet order requirements;

[0025] The calculation formula is:

[0026]

[0027] Where, Indicates order Material requirements;

[0028] The third constraint is that the warehouse’s processing time for an order does not exceed the allowed time;

[0029] The calculation formula is:

[0030]

[0031] Where, Indicates the total time allowed for the warehouse to process all orders;

[0032] The fourth constraint is: the load rate of warehouse handling equipment does not exceed the allowed value;

[0033] The calculation formula is:

[0034]

[0035] Where, Indicates order Delivered to warehouse The load rate increment of the handling equipment during outbound delivery; Represents a warehouse The maximum load rate allowed for handling equipment.

[0036] Furthermore, the allocation execution module receives the allocation results of the solution solving module and generates specific outbound tasks; then it dispatches the warehouse control system WCS of each warehouse to command the logistics access equipment to perform outbound operations; finally, it monitors the execution status of the outbound tasks to ensure that the materials are delivered to the production machines on time.

[0037] Furthermore, the monitoring and prediction module collects order execution data and warehouse status data in real time, including inventory levels, equipment load rates, and order completion status; the monitoring and prediction module learns and predicts historical and real-time order data based on the long short-term memory network LSTM, and predicts the types, quantities, and time of future orders; based on the prediction results, the multi-objective optimization model parameters are adjusted to optimize the allocation strategy.

[0038] Furthermore, the monitoring and prediction module learns and predicts historical and real-time order data based on the long short-term memory network (LSTM) to predict the type, quantity, and time of future orders. The prediction steps are as follows:

[0039] Step 1: Forget gate: decides which information needs to be discarded from the previous state;

[0040] The calculation formula is:

[0041]

[0042] Where, is the current time step The forget gate output of , the value range is [0,1]; is the Sigmoid function, which is used to limit the output between 0 and 1; is the weight matrix of the forget gate, which is used to control the importance of the input; For the previous moment The hidden state and the current input The concatenated vector; Indicates the current time step The input data is a vector containing the feature information of the current moment; is the bias term of the forget gate, used to adjust the output;

[0043] Step 2 Input gate: decides which new information needs to be stored in the cell state and generates candidate cell states;

[0044] The calculation formula is:

[0045] Input gate output:

[0046] Candidate cell states:

[0047] Where, is the current time step The input gate output, with a value range of [0,1], determines which new information will be written into the cell state; is the candidate cell state, which is determined by the hyperbolic tangent function Generate, representing possible new information; 、 are the weight matrices of the input gate and candidate cell states respectively; 、 are the bias terms for the input gate and candidate cell state respectively; in Indicates that the weight matrix belongs to the input gate, to distinguish it from the weight matrices of other gates; in Indicates that the bias term belongs to the input gate, to distinguish it from the bias terms of other gates;

[0048] The cell state is the core of LSTM, which stores long-term dependency information. Its update formula is:

[0049]

[0050] Where, Indicates the current time step The cell state; Represents the previous time step The cell state;

[0051] Step 3 Output gate: determines which information needs to be output from the cell state to the hidden state;

[0052] The calculation formula is:

[0053] Output gate output:

[0054] Hide status update:

[0055] Where, is the current time step The output gate output has a value range of [0,1]; is the weight matrix of the output gate; is the bias term of the output gate; is the current time step The hidden state of is the final output of LSTM;

[0056] Manually adjust the parameters in the multi-objective optimization model based on the predicted future order data , optimize order allocation results.

[0057] A method for allocating orders for production raw materials supplied from multiple warehouses is provided. The method is based on the above-mentioned order allocation system for production raw materials supplied from multiple warehouses and includes the following steps:

[0058] Step S1: Data collection and analysis: The data analysis module extracts data such as production plans and warehouse information, performs abnormality identification and standardization processing, and generates production raw material demand orders;

[0059] Step S2 constructs a multi-objective optimization model and solves it: the solution solving module uses an optimization algorithm to solve the optimal shipping warehouse and quantity for each order based on the comprehensive cost minimization goal and constraints;

[0060] Step S3: Assign tasks and execute outbound delivery: The assignment execution module dispatches logistics equipment in each warehouse based on the solution results, completes the outbound delivery of raw materials and delivers them to the production machines;

[0061] Step S4: Real-time monitoring and order forecasting: The monitoring and forecasting module collects order execution and warehouse status data in real time, uses the LSTM network to predict future order demand, and optimizes the allocation strategy;

[0062] Step S5: Human-computer interaction and optimization: The user interaction module provides an intuitive operation interface, supports manual adjustment of model parameters and exception handling, and further optimizes order allocation results.

[0063] The beneficial effects of the present invention are:

[0064] 1. Minimize overall costs: The system uses a multi-objective optimization model to comprehensively consider factors such as inventory levels, material delivery times, and handling equipment load rates to minimize the overall cost of raw material supply. Compared to traditional single-rule allocation methods, this significantly reduces logistics and warehousing costs.

[0065] 2. Balanced resource utilization: By constraining the load rate of warehouse handling equipment, the system can reasonably allocate tasks, avoiding situations where some warehouses are overloaded while others are idle, improving the utilization rate of equipment resources and extending the service life of equipment.

[0066] 3. Dynamic logistics scheduling: The dynamic adaptability enhancement system combines real-time monitoring and prediction functions to dynamically adjust the allocation strategy according to changes in production order requirements and warehouse status, adapt to complex and changing production environments, and ensure the flexibility and efficiency of logistics scheduling.

[0067] 4. Improved predictive capabilities: By using a long short-term memory (LSTM) network to learn and predict historical and real-time order data, the system can predict future order demand in advance, optimize distribution strategies, reduce logistics pressure during peak periods, and avoid resource waste.

[0068] 5. Human-machine collaborative optimization: The user interaction module supports manual intervention and parameter adjustment, enabling the system to combine human experience with algorithm optimization to further improve the accuracy and practicality of allocation results.

[0069] 6. Efficient JIT supply: Through intelligent scheduling and real-time execution monitoring, the system ensures that production raw materials are delivered to production machines in a JIT manner, reducing inventory backlogs and material waiting time, and improving the operating efficiency of the production line.

[0070] 7. Strong scalability: The system's modular design enables it to have good scalability and can be flexibly adjusted according to the scale and needs of the enterprise. It is suitable for multi-warehouse collaborative operations in different industries and scenarios.

[0071] In summary, this order distribution system for multi-warehouse supply of production raw materials has significant advantages in reducing costs, improving efficiency, and optimizing resource allocation, providing strong technical support for production and logistics collaboration in intelligent manufacturing scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0073] Figure 1 This is a connection diagram of the order distribution system for the multi-warehouse supply of raw materials produced in the present invention;

[0074] Figure 2 This is a flow chart of the order allocation method for multi-warehouse supply of raw materials produced by the present invention. DETAILED DESCRIPTION

[0075] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0076] In order to solve the order allocation problem in multi-warehouse collaborative operations in intelligent manufacturing scenarios, this embodiment provides an order allocation system for multi-warehouse supply of production raw materials; this order allocation system for multi-warehouse supply of production raw materials can comprehensively consider inventory levels, material delivery time, and handling equipment in real time, and monitor the execution of order allocation in real time. Based on real-time and historical order data, it predicts future order demand, manually adjusts multi-objective optimization model parameters, and optimizes order allocation results to achieve efficient and economical JIT supply of production raw materials in intelligent manufacturing scenarios.

[0077] Specifically, such as Figure 1As shown, the order distribution system for multi-warehouse supply of production raw materials includes a data analysis module, a solution solving module, a distribution execution module and a monitoring and prediction module connected with sequential data transmission, and a user interaction module connected with data analysis module, solution solving module, distribution execution module and monitoring and prediction module.

[0078] The data analysis module is responsible for extracting, cleaning, standardizing, and calculating data to generate production raw material demand orders. The data analysis module's hardware includes servers for storing and processing large amounts of data; network equipment such as switches and routers to ensure stable and efficient data transmission; and data acquisition terminals that connect to systems such as ERP, MES, and WMS to obtain real-time data such as production plans and warehouse information. The data analysis module features data acquisition interfaces that connect to ERP, MES, and WMS systems to extract data such as production plans and inventory information. It also includes data cleaning tools for identifying abnormal data and correcting or retrieving it. It also includes a data standardization program that unifies the format and units of data from different sources, as well as a data calculation engine that calculates standardized data according to preset rules to generate production raw material demand orders.

[0079] The solution-solving module uses a multi-objective optimization model to determine the optimal shipping warehouse and quantity. The solution-solving module's hardware includes a high-performance computing server that supports the rapid solution of complex optimization algorithms; a storage device for storing the parameters, constraints, and solution results of the multi-objective optimization model. The solution-solving module is equipped with an optimization algorithm library, including linear programming and mixed integer programming, for solving multi-objective optimization problems; a modeling tool for defining objective functions and constraints and constructing comprehensive cost minimization models; and a scheduling engine that generates specific warehouse allocation plans based on the optimization results.

[0080] The allocation execution module allocates the same raw materials in the order to different warehouses based on the results of the solution solving module, and dispatches logistics equipment to complete the outbound operation. The hardware equipment of the allocation execution module includes: a warehouse control system WCS that directs stackers, conveyors and other logistics storage and retrieval equipment to perform tasks, stackers, automatic guided vehicles AGV, conveyor belts and other logistics storage and retrieval equipment used for material storage and transportation, RFID barcode scanning equipment used to identify material specifications and track logistics status, and a monitoring screen that displays the operating status of warehouse equipment and task progress. The allocation execution module is equipped with a task scheduling program that generates specific outbound tasks based on the optimization results and assigns them to the corresponding equipment; an equipment control program that interacts with the WCS system to control the operation of logistics equipment; and a status monitoring program that collects equipment operating status and task completion status in real time.

[0081] The monitoring and prediction module monitors order fulfillment and warehouse status in real time, predicts future order demand based on historical and real-time data, and optimizes allocation strategies. The hardware components of this module include a monitoring terminal that displays information such as order fulfillment progress and warehouse status; inventory and load sensors that collect real-time warehouse data; and big data storage for historical order data and real-time monitoring data. The module includes a built-in real-time monitoring program to collect and display order fulfillment and warehouse status; a prediction model, namely an LSTM network, to predict future order demand based on historical and real-time data; and an optimization program that dynamically adjusts the parameters of the multi-objective optimization model based on the prediction results.

[0082] The user interaction module provides a human-computer interface, supporting manual intervention and parameter adjustment to optimize order allocation results. The hardware for the user interaction module includes a PC, tablet, or touchscreen for user operation and viewing system information, as well as an audible and visual alarm to alert users to abnormal situations. The user interaction module features a graphical user interface (GUI) that visually displays information such as order allocation results, warehouse status, and logistics execution. It also includes a parameter adjustment tool that allows users to manually adjust key parameters such as weight coefficients and constraints. Furthermore, it features an exception alarm program that issues alerts and logs abnormalities.

[0083] Furthermore, each module in this embodiment is described in detail below:

[0084] The data analysis module is the core starting point of the entire production raw material multi-warehouse supply order distribution system. It is responsible for extracting, cleaning, standardizing, and calculating production raw material demand orders from raw data. The function implementation of this module is divided into the following key steps:

[0085] First, extract the data:

[0086] Data extraction can be automatically triggered based on a set time point, or it can be manually initiated based on actual needs. The data comes from predetermined data sources, including but not limited to the following systems: ERP system provides information such as production plans, product BOMs, material requirements, etc.; MES system provides real-time information such as production equipment status and production progress; WMS system provides information such as warehouse inventory, material specifications, storage locations, etc.; logistics management system provides logistics-related information such as delivery time and handling equipment load rate. Extract a variety of data related to the supply of production raw materials, including: clarifying the production time and production quantity of the product, that is, the production plan; listing the types and quantities of raw materials required for each product, that is, the product BOM; inventory, material specifications, storage location, packaging information, etc. of each warehouse, that is, warehouse information; material delivery time, handling equipment load rate, etc., that is, logistics information.

[0087] Then, perform anomaly identification:

[0088] To ensure data quality and reliability and avoid biases in subsequent analysis and decision-making due to abnormal data, anomaly identification is performed. This includes checking data integrity, such as missing values ​​or incomplete fields; checking data rationality, such as whether inventory levels are negative or production quantities exceed reasonable ranges; and checking data consistency, such as whether linked data across different data sources is consistent. The system flags abnormal data and attempts to retrieve the correct data. If the correct data cannot be retrieved, the system logs the anomaly and notifies relevant personnel for manual intervention.

[0089] Then, perform the normalization transformation:

[0090] For normal data, data from different sources, formats, and units is converted into a unified standard format to facilitate subsequent processing and analysis. Standardization conversion includes: standardizing data formats, such as standardizing dates to "YYYY-MM-DD" and retaining values ​​to two decimal places; standardizing measurement units, such as standardizing inventory units to "tons" or "pieces"; and data code mapping, such as mapping material specification names to unique codes for easier system recognition and processing. After standardization conversion, a standardized data set is output, with consistent formats and units across all fields, facilitating subsequent structural conversion and calculations.

[0091] Then, perform the structure conversion:

[0092] Standardized data is organized according to pre-set logical relationships to form a structured data model, facilitating subsequent calculations. Structural transformation includes: data grouping, for example, grouping all relevant data for the same order into a single data structure; data association, for example, linking production plans, bills of materials, and inventory information through material coding; and data aggregation, for example, aggregating inventory data from multiple warehouses by material type. Structural transformation outputs a structured dataset with clear hierarchical relationships and logical associations.

[0093] Then, perform data calculation:

[0094] Based on pre-set data correspondences, structured data is calculated to generate production raw material demand orders. This data calculation includes: calculating the demand for each material based on the production plan and bill of materials (BOM); determining the material's delivery destination based on the production equipment's location information (delivery destination calculation); and calculating the time required for the material to be shipped from the warehouse to the production machine (estimated delivery time) based on logistics information. The above data calculation process is conventional in this field, and the calculation formula is not detailed here. The output of this data calculation is a production raw material demand order. Each order contains: a unique identifier (order number); the type of raw material to be delivered (material type); the quantity of raw material to be delivered (demand quantity); the location of the production machine to be delivered (delivery destination); and the time the material is expected to arrive at the production machine (estimated delivery time). Production raw material demand orders are core data; each order corresponds to a raw material, which may be shipped from one or more warehouses and ultimately supplied to a production unit.

[0095] Finally, output and transfer:

[0096] The data analysis module ultimately outputs the production raw material demand order as structured data, which is then passed to the solution-solving module. This data can be delivered to the solution-solving module via an API or other data transmission protocol. The module also records key steps and exceptions in the order generation process for easy traceability.

[0097] The data analysis module generates high-quality production raw material demand orders by extracting, cleaning, standardizing, structuring, and calculating raw data. Its core role is to provide an accurate and reliable data foundation for subsequent order allocation and optimization, thereby ensuring the efficient operation of the entire system and the accuracy of decision-making.

[0098] The solution-solving module is the core of the entire multi-warehouse supply and order allocation system for production raw materials. It is responsible for calculating the optimal dispatch warehouse and quantity allocation plan through a mathematical multi-objective optimization model. Aiming to minimize the overall cost of production raw material supply, the solution-solving module defines four constraints: the first constraint, the second constraint, the third constraint, and the fourth constraint. The solution-solving module uses the production raw material demand order data generated by the data analysis module as input and solves the multi-objective optimization model to determine the dispatch warehouse and corresponding quantity for each production raw material.

[0099] First, the comprehensive cost is minimized. The calculation formula of the multi-objective optimization model is:

[0100]

[0101] Where, It represents the comprehensive cost of supplying raw materials from multiple warehouses; Indicates the total number of orders, that is, the number of production raw material demand orders that need to be allocated; Indicates the total number of warehouses, that is, the number of warehouses available for selection; Indicates the index of the order, , used to identify the orders; Represents the index of the warehouse, , used to identify the warehouses; Represents a decision variable, which takes a value of 0 or 1. Indicates order By warehouse deal with, Indicates order Not by warehouse deal with; Represents the weight coefficients of different cost factors, satisfying , is the weight of inventory-related costs, is the weight of the material transportation time-related cost, The weight of the costs associated with the loading rate of the handling equipment; Indicates order Materials in the warehouse Inventory in Indicates order Materials in the warehouse The time required to ship from the warehouse to the destination; Represents a warehouse Loading rate of handling equipment; It means to perform a double summation on all orders and warehouses, that is, to comprehensively consider the distribution relationship between each order and each warehouse.

[0102] By using the comprehensive cost of raw material supply Minimum is the goal. Comprehensively consider inventory cost , transportation time cost , handling equipment load rate cost There are three cost factors. The inventory cost weight is , the larger the inventory, the lower the cost of shipping from the warehouse. The transportation time cost weight is The longer the transportation time, the lower the logistics efficiency and the higher the cost. The weight of the handling equipment load rate cost is The higher the load rate of the handling equipment, the tighter the utilization of equipment resources and the higher the cost. satisfy , used to balance the importance of different cost factors.

[0103] To ensure the feasibility of the allocation plan, the four constraints of the solution solving module are as follows:

[0104] The first constraint is: each order can only be processed by one warehouse;

[0105] The calculation formula is:

[0106]

[0107] By the first constraint, each order Must and can only be assigned to one warehouse , thus avoiding orders being duplicated or missed.

[0108] The second constraint is: the warehouse inventory must meet order requirements;

[0109] The calculation formula is:

[0110]

[0111] Where, Indicates order Material requirements;

[0112] By the second constraint, if the order Assigned to warehouse ,Right now , then the warehouse Inventory Must be greater than or equal to order Material requirements , thereby ensuring that the warehouse has sufficient inventory to meet order requirements.

[0113] The third constraint is that the warehouse’s processing time for an order does not exceed the allowed time;

[0114] The calculation formula is:

[0115]

[0116] Where, Indicates the total time allowed for the warehouse to process all orders;

[0117] By the third constraint, the sum of the delivery time of all orders cannot exceed the total allowed time , thereby ensuring the timeliness of logistics distribution and avoiding the impact of timeouts on production progress.

[0118] The fourth constraint is: the load rate of warehouse handling equipment does not exceed the allowed value;

[0119] The calculation formula is:

[0120]

[0121] Where, Indicates order Delivered to warehouse The load rate increment of the handling equipment during outbound delivery; Represents a warehouse The maximum load rate allowed for handling equipment;

[0122] By the fourth constraint, the warehouse The total load rate of the handling equipment cannot exceed the maximum load rate allowed , avoid equipment overload and ensure the normal operation and service life of the equipment.

[0123] The solution-solving module inputs: production raw material demand order data and warehouse-related information. The production raw material demand order data, obtained from the data analysis module, includes information such as material type, demand quantity, and delivery destination for each order. Warehouse-related information includes inventory levels, delivery times, and handling equipment load rates. The solution-solving module outputs: outbound warehouses and quantities, as well as an allocation plan. The outbound warehouses and quantities are determined for each order based on the results of the multi-objective optimization model. The allocation plan generates a specific warehouse allocation plan, which serves as input to the allocation execution module.

[0124] The solution-solving module generates an optimal order allocation plan by comprehensively considering factors such as inventory levels, delivery times, and equipment load rates, while also combining them with strict constraints. The design and implementation of this module significantly enhances the intelligent supply of raw materials across multiple warehouses.

[0125] The distribution execution module is a key component of the multi-warehouse supply order distribution system for production raw materials. It is responsible for translating the optimal distribution results generated by the solution solution module into actual logistics operations. By scheduling the warehouse control system (WCS) and logistics equipment, this module completes the outbound delivery of production raw materials and delivers them to production machines on time.

[0126] Based on the optimization results of the solution solving module, the allocation execution module allocates the same raw materials in the production raw material demand order to different warehouses for outbound operations, and ensures that the materials are delivered to the production machines on time, thereby achieving efficient and accurate logistics scheduling and ensuring the JIT supply of production raw materials.

[0127] The input data of the allocation execution module includes: optimized allocation results, warehouse status information, and logistics path information; among them, the optimized allocation results are the shipping warehouses and corresponding shipping quantities corresponding to each order obtained from the solution solution module; warehouse status information includes real-time data such as the inventory level of each warehouse, equipment load rate, material storage location, etc.; logistics path information includes material transportation path, estimated transportation time, etc.

[0128] The execution flow of the allocation execution module is as follows:

[0129] First, generate the outbound task:

[0130] According to the optimization allocation results, each order is broken down into specific outbound tasks. For example: Order Materials from the warehouse Outbound Unit, by warehouse Outbound Unit. Deliver outbound tasks to the warehouse's WCS system in a standardized format, including information such as material type, outbound quantity, storage location, and target production machine.

[0131] Then, dispatch logistics access equipment:

[0132] Dispatching logistics storage and retrieval equipment, such as stackers, automated guided vehicles (AGVs), and conveyor belts, to execute outbound tasks. Based on task instructions, the WCS controls the logistics storage and retrieval equipment to complete the following operations: Locating materials: locating the target material based on storage location information; extracting materials: using stackers or AGVs to remove materials from storage; and loading materials: loading materials onto conveyor equipment.

[0133] Then, dispatch logistics transportation equipment:

[0134] Based on logistics routing information, we plan the optimal material delivery route from the warehouse to the production machine. Logistics transportation equipment, such as conveyor belts and elevators, transport materials to the target production machine and monitor the delivery status in real time to ensure timely delivery of materials.

[0135] Finally, task completion feedback:

[0136] After the task is completed, the warehouse inventory, equipment status and other information will be updated. If any abnormalities occur during the execution process, such as equipment failure or path blockage, an alarm will be issued in time and the emergency plan will be triggered.

[0137] Through the allocation execution module, materials are shipped from designated warehouses and successfully delivered to production machines. Simultaneously, data such as warehouse inventory, equipment status, and task completion status are updated to provide a basis for subsequent scheduling. By scheduling logistics access equipment and conveying equipment, the allocation execution module translates optimized allocation results into specific outbound operations, ensuring the efficient supply of raw materials for production. The design and implementation of this module not only enhances the automation and intelligence of logistics execution but also provides strong support for production and logistics collaboration in intelligent manufacturing scenarios.

[0138] The monitoring and prediction module is a crucial component of the multi-warehouse supply and order distribution system for production raw materials. It monitors order execution and warehouse status in real time. Using long-short-term memory (LSTM) networks, the module uses big data learning and optimization based on historical and real-time data to predict future order demand. This prediction results in turn help refine the multi-objective optimization model, improving the system's intelligence and resource utilization efficiency.

[0139] The monitoring and prediction module monitors the execution of production raw material orders and warehouse status in real time. Based on real-time and historical order data, it uses an LSTM network to predict production raw material demand. Based on the prediction results, it manually adjusts the multi-objective optimization model parameters and optimizes the allocation strategy. This improves the foresight and flexibility of order allocation, ensures that the system can dynamically adapt to future changes in production demand, reduces peak logistics pressure, avoids resource waste, and improves overall supply chain efficiency.

[0140] The input data for the monitoring and forecasting module includes real-time data, historical data, and external data. Real-time data includes information such as the current order execution progress, warehouse inventory, equipment load rate, and logistics transportation status; historical data includes information such as order type, quantity, and time distribution over a period of time; and external data includes production plans, market trends, and seasonal demand fluctuations, which are used to assist in forecasting.

[0141] LSTM is a special type of recurrent neural network (RNN) that can capture long-term dependencies in time series data. It updates cell states and selects information through three key steps: the forget gate, the input gate, and the output gate. The monitoring and prediction module uses the long short-term memory (LSTM) network to learn and predict historical and real-time order data, predicting the type, quantity, and timing of future orders. The prediction steps are as follows:

[0142] Step 1: Forget gate: decides which information needs to be discarded from the previous state;

[0143] The calculation formula is:

[0144]

[0145] Where, is the current time step The forget gate output of , the value range is [0,1]; is the Sigmoid function, which is used to limit the output between 0 and 1; is the weight matrix of the forget gate, which is used to control the importance of the input; For the previous moment The hidden state and the current input The concatenated vector; Indicates the current time step The input data is a vector containing the feature information of the current moment; is the bias term of the forget gate, which is used to adjust the output.

[0146] Step 2 Input gate: decides which new information needs to be stored in the cell state and generates candidate cell states;

[0147] The calculation formula is:

[0148] Input gate output:

[0149] Candidate cell states:

[0150] Where, is the current time step The input gate output, with a value range of [0,1], determines which new information will be written into the cell state; is the candidate cell state, which is determined by the hyperbolic tangent function Generate, representing possible new information; 、 are the weight matrices of the input gate and candidate cell states respectively; 、 are the bias terms for the input gate and candidate cell state respectively; in Indicates that the weight matrix belongs to the input gate, to distinguish it from the weight matrices of other gates; in Indicates that the bias term belongs to the input gate, to distinguish it from the bias terms of other gates.

[0151] The cell state is the core of LSTM, which stores long-term dependency information. Its update formula is:

[0152]

[0153] Where, Indicates the current time step The cell state; Represents the previous time step cell state.

[0154] Step 3 Output gate: determines which information needs to be output from the cell state to the hidden state;

[0155] The calculation formula is:

[0156] Output gate output:

[0157] Hide status update:

[0158] Where, is the current time step The output gate output has a value range of [0,1]; is the weight matrix of the output gate; is the bias term of the output gate; is the current time step The hidden state of is used as the final output of LSTM.

[0159] The LSTM network predicts future order demands by learning from historical and real-time data, including order types, order quantities, and order times. The monitoring and prediction module generates forecast data for future order demands as a basis for optimizing the multi-objective optimization model; based on the forecast results, the parameters in the multi-objective optimization model are manually adjusted. , to optimize the allocation strategy. For example: increase the weight of inventory Prioritize warehouses with sufficient inventory; increase the weight of delivery time To optimize logistics efficiency; increase load rate weight To balance device resources.

[0160] The monitoring and prediction module uses real-time monitoring and LSTM network prediction to accurately predict and dynamically optimize future order demand. This module not only enhances the intelligence of order allocation but also provides strong technical support for production and logistics collaboration in intelligent manufacturing scenarios.

[0161] The user interaction module is a crucial component of the multi-warehouse supply order distribution system for production raw materials. It provides a human-computer interface and supports real-time interaction between users and the system. By connecting the data analysis module, solution-solving module, distribution execution module, and monitoring and prediction module, this module transmits human-computer interaction information, enabling users to monitor the system's operating status in real time and perform manual intervention and optimization adjustments when necessary.

[0162] The user interaction module provides an intuitive human-computer interaction interface, displays key information about system operation, supports users to manually adjust multi-objective optimization model parameters, constraints, etc., and provides real-time feedback on system operation status, including order execution progress, warehouse status, logistics equipment operation status, etc. It issues alarms in abnormal situations and records abnormal logs to enhance the transparency and controllability of the system, ensure that users can flexibly adjust system behavior according to actual conditions, and provide decision support tools to help users optimize order allocation strategies.

[0163] Input data for the user interaction module includes system status information, forecast results, and user input. System status information includes order fulfillment progress, warehouse inventory levels, equipment load rates, and logistics delivery status. Forecast results come from the future order demand forecast data from the monitoring and forecasting module. User input includes adjustment instructions or optimization suggestions entered through the user interface.

[0164] The user interaction module displays information such as the allocation of production raw material demand orders, including the shipping warehouse, quantity, and delivery time; the real-time status of each warehouse, including inventory levels, equipment load rate, and task execution progress; the operating status of logistics equipment, such as stackers, AGVs, and conveyor belts; and forecasts, such as the type, quantity, and time distribution of future orders. This information can be intuitively displayed through charts, tables, and maps, dynamically displaying key indicators such as inventory levels and equipment load rates. When an abnormality is detected, the user is alerted to the situation with audio and visual prompts or pop-up windows. Ultimately, the user interaction module generates a new order allocation plan based on the user's adjustment instructions, updating information such as order execution progress, warehouse status, and logistics equipment operation in real time. Once the user confirms and resolves the abnormality, the system resumes normal operation.

[0165] The user interaction module provides an intuitive graphical interface and powerful interactive features, enabling users to monitor system operating status in real time and perform manual intervention and optimization adjustments when necessary. The design and implementation of this module not only enhances the system's usability and flexibility but also provides reliable technical support for production and logistics collaboration in intelligent manufacturing scenarios.

[0166] Based on the above-mentioned order allocation system for multi-warehouse supply of production raw materials, this embodiment also proposes an order allocation method for multi-warehouse supply of production raw materials; Figure 2 As shown, the order allocation method for multi-warehouse supply of production raw materials includes the following steps:

[0167] Step S1: Data collection and analysis: The data analysis module extracts data such as production plans and warehouse information, performs abnormality identification and standardization processing, and generates production raw material demand orders;

[0168] Step S2 constructs a multi-objective optimization model and solves it: the solution solving module uses an optimization algorithm to solve the optimal shipping warehouse and quantity for each order based on the comprehensive cost minimization goal and constraints;

[0169] Step S3: Assign tasks and execute outbound delivery: The assignment execution module dispatches logistics equipment in each warehouse based on the solution results, completes the outbound delivery of raw materials and delivers them to the production machines;

[0170] Step S4: Real-time monitoring and order forecasting: The monitoring and forecasting module collects order execution and warehouse status data in real time, uses the LSTM network to predict future order demand, and optimizes the allocation strategy;

[0171] Step S5: Human-computer interaction and optimization: The user interaction module provides an intuitive operation interface, supports manual adjustment of model parameters and exception handling, and further optimizes order allocation results.

[0172] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An order distribution system for multi-warehouse supply of production raw materials, characterized by: The order distribution system for multi-warehouse supply of production raw materials includes a data analysis module, a solution solving module, a distribution execution module and a monitoring and prediction module connected in sequence data transmission, and a user interaction module connected in data transmission with the data analysis module, the solution solving module, the distribution execution module and the monitoring and prediction module; The data analysis module is responsible for extracting, cleaning, standardizing, and calculating data to generate production raw material demand orders. The solution-solving module uses a multi-objective optimization model to determine the optimal shipping warehouse and quantity. The allocation execution module allocates the same raw materials in the order to different warehouses based on the solution-solving module's results and dispatches logistics equipment to complete the shipping operation. The monitoring and prediction module monitors order execution and warehouse status in real time, predicts future order demand based on historical and real-time data, and optimizes allocation strategies. The user interaction module provides a human-computer interaction interface, supports manual intervention and parameter adjustment, and optimizes order allocation results; The solution solving module aims to minimize the comprehensive cost of raw material supply and defines four constraints: The first constraint is: each order can only be processed by one warehouse; The second constraint is: the warehouse inventory must meet order requirements; The third constraint is that the warehouse’s processing time for an order does not exceed the allowed time; The fourth constraint is: the load rate of warehouse handling equipment does not exceed the allowed value; When each order corresponds to a type of raw material, the solution solving module takes the production raw material demand order data generated by the data analysis module as input and determines the outbound warehouse and corresponding quantity of each production raw material by solving the multi-objective optimization model; The calculation formula of the multi-objective optimization model is: ； Where, It represents the comprehensive cost of supplying raw materials from multiple warehouses; Indicates the total number of orders, that is, the number of production raw material demand orders that need to be allocated; Indicates the total number of warehouses, that is, the number of warehouses available for selection; Indicates the index of the order, , used to identify the orders; Represents the index of the warehouse, , used to identify the warehouses; Represents a decision variable, which takes a value of 0 or 1. Indicates order By warehouse deal with, Indicates order Not by warehouse deal with; Represents the weight coefficients of different cost factors, satisfying , is the weight of inventory-related costs, is the weight of the material transportation time-related cost, The weight of the costs associated with the loading rate of the handling equipment; Indicates order Materials in the warehouse Inventory in Indicates order Materials in the warehouse The time required to ship from the warehouse to the destination; Represents a warehouse Loading rate of handling equipment; It means to perform a double summation on all orders and warehouses, that is, to comprehensively consider the distribution relationship between each order and each warehouse.

2. The order distribution system for multi-warehouse supply of production raw materials according to claim 1 is characterized by: The data analysis module extracts data from a predetermined data source based on a set time point or manually triggered to generate a production raw material data set; it then identifies anomalies in the extracted data and re-acquires any abnormal data found; For normal data, standardization conversion is performed to unify the data format and unit; then the standardized data is structured and organized according to the preset logical relationship to form a clear data structure; finally, based on the set correspondence between data, the structured data is calculated to generate production raw material demand orders.

3. The order distribution system for multi-warehouse supply of production raw materials according to claim 2 is characterized by: The production raw material data set includes: production plan, product production time, production quantity, product BOM table, warehouse information, and corresponding warehouse material specification information, material quantity information and material packaging information.

4. The order distribution system for multi-warehouse supply of production raw materials according to claim 2 is characterized by: In the order distribution system for multi-warehouse supply of production raw materials, the production raw material demand order is one of the core data; each order corresponds to a raw material, which may be shipped from one or more warehouses and eventually supplied to a production unit for use.

5. The order distribution system for multi-warehouse supply of production raw materials according to claim 1 is characterized by: The calculation formula for the first constraint is: ； The calculation formula for the second constraint is: ； Where, Indicates order Material requirements; The calculation formula for the third constraint is: ； Where, Indicates the total time allowed for the warehouse to process all orders; The calculation formula for the fourth constraint is: ； Where, Indicates order Delivered to warehouse The load rate increment of the handling equipment during outbound delivery; Represents a warehouse The maximum load rate allowed for handling equipment.

6. The order distribution system for multi-warehouse supply of production raw materials according to claim 1 is characterized by: The allocation execution module receives the allocation results of the solution solving module and generates specific outbound tasks; then it dispatches the warehouse control system (WCS) of each warehouse to instruct the logistics storage and access equipment to execute the outbound operation; finally, it monitors the execution status of the outbound tasks to ensure that the materials are delivered to the production machines on time.

7. The order distribution system for multi-warehouse supply of production raw materials according to claim 1 is characterized by: The monitoring and prediction module collects order execution data and warehouse status data in real time, including inventory levels, equipment load rates, and order completion status. The monitoring and prediction module learns and predicts historical and real-time order data based on the long short-term memory network (LSTM), predicting the type, quantity, and time of future orders. Based on the prediction results, the multi-objective optimization model parameters are adjusted to optimize the allocation strategy.

8. The order distribution system for multi-warehouse supply of production raw materials according to claim 7 is characterized by: The monitoring and prediction module learns and predicts historical and real-time order data based on the long short-term memory network (LSTM) to predict the type, quantity, and time of future orders. The prediction steps are as follows: Step 1: Forget gate: decides which information needs to be discarded from the previous state; The calculation formula is: ； Where, is the forget gate output of the current time step t, and its value range is [0,1]; is the Sigmoid function, which is used to limit the output between 0 and 1; is the weight matrix of the forget gate, which is used to control the importance of the input; is the hidden state at the previous moment t-1 and the current input The concatenated vector; Represents the input data of the current time step t, which is a vector containing the feature information of the current moment; is the bias term of the forget gate, used to adjust the output; Step 2 Input gate: decides which new information needs to be stored in the cell state and generates candidate cell states; The calculation formula is: Input gate output: ; Candidate cell states: ; Where, is the current time step The input gate output, with a value range of [0,1], determines which new information will be written into the cell state; is the candidate cell state, which is determined by the hyperbolic tangent function Generate, representing possible new information; 、 are the weight matrices of the input gate and candidate cell states respectively; 、 are the bias terms for the input gate and candidate cell state respectively; in Indicates that the weight matrix belongs to the input gate, to distinguish it from the weight matrices of other gates; in Indicates that the bias term belongs to the input gate, to distinguish it from the bias terms of other gates; The cell state is the core of LSTM, which stores long-term dependency information. Its update formula is: ； Where, Indicates the current time step The cell state; Represents the previous time step The cell state; Step 3 Output gate: determines which information needs to be output from the cell state to the hidden state; The calculation formula is: Output gate output: ; Hide status update: ; Where, is the current time step The output gate output has a value range of [0,1]; is the weight matrix of the output gate; is the bias term of the output gate; is the current time step The hidden state of is the final output of LSTM; Manually adjust the parameters in the multi-objective optimization model based on the predicted future order data , optimize order allocation results.

9. A method for allocating orders for production raw materials supplied from multiple warehouses, the method being based on the system for allocating orders for production raw materials supplied from multiple warehouses as claimed in claim 1, characterized in that: The following steps are involved: Step S1: Data collection and analysis: The data analysis module extracts data such as production plans and warehouse information, performs abnormality identification and standardization processing, and generates production raw material demand orders; Step S2 constructs a multi-objective optimization model and solves it: the solution solving module uses an optimization algorithm to solve the optimal shipping warehouse and quantity for each order based on the comprehensive cost minimization goal and constraints; Step S3: Assign tasks and execute outbound delivery: The assignment execution module dispatches logistics equipment in each warehouse based on the solution results, completes the outbound delivery of raw materials and delivers them to the production machines; Step S4: Real-time monitoring and order forecasting: The monitoring and forecasting module collects order execution and warehouse status data in real time, uses the LSTM network to predict future order demand, and optimizes the allocation strategy; Step S5: Human-computer interaction and optimization: The user interaction module provides an intuitive operation interface, supports manual adjustment of model parameters and exception handling, and further optimizes order allocation results.

Images