Distributed warehouse multi-source data centralized management system based on artificial intelligence

Abstract

The invention discloses a distributed warehouse multi-source data centralized management system based on artificial intelligence, relates to the technical field of warehouse management, and solves the problem that in the prior art, when goods are stored and taken, storage and taking congestion is easily caused due to the fact that the storage positions of the goods are not distinguished. According to the invention, a total scheduling unit is arranged to distribute goods, goods produced by a factory or from other goods sources are distributed to each distributed warehouse, and all process data and result data in the process are recorded. Separate storage facilitates storage and taking of goods in different orders on one hand, and facilitates acquisition of two different data on the other hand, so that management personnel can manage the data conveniently. In addition, a terminal and a material management system are further arranged and used for displaying warehouse-in, warehouse-out and stock ledgers and counting indexes such as the material type number of the warehouse, the material stock early warning number, newly-added materials of the current year, the material warehouse-in order number of the current year and the material warehouse-out order number of the current year to master the general situation of the warehouse.

Description

Technical Field

[0001] The present invention relates to the technical field of warehouse management, and in particular to an artificial intelligence-based distributed warehouse multi-source data centralized management system. Background Art

[0002] In today's logistics and supply chain management, warehouse storage has become a core component of the flow of goods, pervasive throughout the storage and distribution of all types of goods. Whether it's raw material storage in the manufacturing industry, temporary product storage in the e-commerce industry, or inventory management in the retail industry, warehouses play an indispensable role. However, with the acceleration of global trade and the booming e-commerce industry, the traditional centralized warehouse storage model is no longer able to meet increasingly complex market demands. Modern trading markets demand greater flexibility and efficiency in the allocation of goods, leading more and more companies to turn to distributed warehouse storage. By deploying warehouses in multiple locations, companies can respond to customer needs more quickly, reduce transportation costs, optimize inventory turnover, and enhance overall supply chain resilience. This decentralized storage strategy not only improves the accessibility of goods but also provides greater resilience to emergencies, making it an inevitable trend in the current market environment.

[0003] Despite the importance of warehouse storage in modern logistics, many warehouses currently lack significant design and management deficiencies, particularly in the area of storage zoning. Many traditional warehouses lack clear demarcations between storage areas, resulting in mixed storage of goods and inefficient management. Even some warehouses have introduced data management systems, which are largely based on undivided storage data, making it difficult to accurately arrange the storage location of goods and achieve efficient inventory management. During order processing, the lack of location-based storage strategies makes the storage and retrieval of goods extremely inconvenient. The lack of differentiation in the storage location of goods during storage and retrieval can easily lead to congestion and other issues. Therefore, optimizing the demarcation of warehouse storage areas and introducing more refined data management have become key to improving warehouse efficiency and responsiveness.

[0004] In view of this, an artificial intelligence-based distributed warehouse multi-source data centralized management system is needed. Summary of the Invention

[0005] To address the problem in existing technologies where access to goods is often congested due to the lack of differentiation in the storage location of goods, the present invention provides an AI-based distributed warehouse multi-source data centralized management system that can perform classified storage of goods. This not only facilitates the storage and retrieval of goods for different orders, maximizing customer satisfaction and allowing as many customers as possible to reduce the time it takes to retrieve goods from the warehouse, but also facilitates the acquisition of two different types of data for management personnel. The specific technical solution is as follows:

[0006] An artificial intelligence-based distributed warehouse multi-source data centralized management system includes at least one general scheduling unit and at least two distributed warehouse units and execution units, wherein:

[0007] The general dispatching unit is used to obtain user orders and the storage status of each distributed warehouse unit, and instruct the factory source to distribute and deliver goods to each distributed warehouse, or allocate goods between distributed warehouse units. In addition, the general dispatching unit also provides the specific storage location of the goods based on the different pick-up times of user scheduled orders;

[0008] The distributed warehouse unit is used to store goods. The distributed warehouse unit is divided into a predetermined storage area and a basic storage area. The predetermined storage area is used to store goods required for user predetermined orders in a sequential order based on the pick-up time and the ordered goods. The basic storage area is used to store goods directly allocated by the general scheduling unit.

[0009] The execution unit is used to receive instructions from the general scheduling unit and transport the goods to the specific storage location in the distributed warehouse unit.

[0010] Preferably, the general dispatching unit distributes the goods, distributes the goods produced by the factory or from other sources to each distributed warehouse, and records all process data and result data in the process. The process data includes user orders and goods allocation plans, and the result data is the outbound goods data, inbound goods data and goods data currently stored in the warehouse of each distributed warehouse.

[0011] Preferably, in the distributed warehouse, there are two types of cargo storage areas, one is the basic storage area and the other is the reserved storage area; wherein, the basic storage area is used to store cargo that is directly delivered from the factory or other cargo sources, and is used for direct acquisition by customers with real-time orders, and the reserved storage area is used to store cargo required based on user reserved orders, and each cargo has its fixed output object and output time.

[0012] Preferably, the overall dispatching unit includes a data interaction module, a data storage module, a warehouse matching module and an alarm module;

[0013] The data interaction module is used to obtain real-time customer orders and pre-ordered customer orders, and output the final transportation plan to the execution unit; wherein the information in the real-time customer orders includes at least customer information and order information, the customer information includes at least the customer address, and the order information includes at least the type and quantity of goods; in addition to the customer information and order information, the pre-ordered customer orders also include the expected pickup time;

[0014] The data storage module stores the address of each distributed warehouse unit;

[0015] The warehouse matching module selects a specific distributed warehouse unit for storing the goods corresponding to the customer's scheduled order based on the address of the customer's scheduled order;

[0016] The alarm module is used to send an alarm message to the manager terminal when the warehouse cannot select a specific distributed warehouse unit.

[0017] Preferably, a selection function is provided in the warehouse matching module, and the specific storage warehouse of the goods is obtained by solving the selection function. The specific selection function is as follows:

[0018]

[0019] Where Z is the matching function, S 间距 is the distance between the customer address and the distributed warehouse unit address, R 预定区_剩余 The remaining storage capacity of the reserved storage area in the distributed warehouse unit.

[0020] For the matching function, there are residual capacity constraints and spacing constraints to avoid the problem that the residual capacity of distributed warehouse units is insufficient to accommodate goods and the spacing is too large.

[0021] The remaining capacity constraints are as follows:

[0022] R 剩余 ≥R 货物

[0023] Where R 货物 The capacity of the cargo;

[0024] The spacing constraints are as follows:

[0025] S 间距 ≤S 间距_max

[0026] Where S 间距_max is the maximum tolerable spacing.

[0027] Preferably, the redundancy factor is considered in the remaining capacity constraint, as follows:

[0028] R 剩余 ≥R 货物 +δ

[0029] Where R 货物 is the capacity of the cargo, δ is the redundancy coefficient, which is set as a constant and has a value range of δ≥0;

[0030] When selecting a specific distributed warehouse, the specific information of each warehouse is substituted into the matching function for solution to obtain the optimal distributed warehouse unit for storage; in this process, if there is no solution, the value of the redundancy coefficient δ is further reduced and the solution is performed again. If the redundancy coefficient δ has dropped to 0 and there is still no solution, the alarm module is activated and an alarm message is sent to the administrator terminal.

[0031] Preferably, the general scheduling unit further includes a location matching module, which is used to select a specific location in the distributed warehouse. In the distributed warehouse unit, a predetermined storage area is provided with a plurality of storage units, and the storage units are classified according to different storage capacities. Each storage unit under each category has a unique number, and there is a difference in order between the plurality of storage units under each category.

[0032] The specific storage location of the reserved storage area is determined as follows:

[0033] S1: Determine the storage department category for storing the goods according to the capacity of the goods. After the storage department category is determined, select a storage department from several storage departments under this storage department category to store the goods.

[0034] S2: Get the pick-up time of the scheduled order to be assigned, and determine whether there are goods stored in the storage unit of the current category. If so, get the specific pick-up time and placement location of the goods with the smallest difference from the pick-up time of the scheduled order to be assigned as the reference time and reference location, and then proceed to step S3; if not, store the goods of the scheduled order in the storage unit numbered x. 2 / n (If n is an odd number, 2 / n is an integer), that is, it is stored in the middle position of the sequence;

[0035] S3: Determine whether the pickup time of the scheduled order to be allocated is before or after the reference time. If it is before the reference time, the goods position of the scheduled order to be allocated that is in the first N positions of the reference position sequence is used as the position to be allocated. Otherwise, the goods position of the scheduled order to be allocated that is in the last N positions of the reference position sequence is used as the position to be allocated. The specific value of N is as follows:

[0036]

[0037] Where, T 待分配 is the pickup time of the order to be assigned, T 参考 is the reference time, T 间隔 is the average interval time of the storage department.

[0038] Average storage interval time T 间隔 Determined by:

[0039]

[0040] Where, T max is the delivery time with the largest difference from the current time in the current storage department, T min N is the minimum delivery time in the current storage department and the current time. max-min T max and T min The interval between the specific storage locations of the goods of the corresponding order, that is, how many storage units are left empty in between.

[0041] Preferably, a deep learning model is trained to predict the order pick-up time and pick-up quantity at future moments, and the order pick-up time and pick-up quantity at future moments are sent to the manager terminal for decision-making reference, including planning various types of storage space in each predetermined storage unit.

[0042] Preferably, it further includes a material management unit, which includes two components: a terminal and a material management system, wherein:

[0043] The terminal is used to display incoming and outgoing warehouse information and inventory records, and to compile statistics on the number of material types in the warehouse, the number of material inventory warnings, the number of new materials added this year, the number of incoming material orders this year, and the number of outgoing material orders this year to understand the warehouse situation. In addition, facial recognition and facial recognition training are also performed.

[0044] The material management system displays the distribution of all materials under the jurisdiction of the unit based on the Zhikan map, counts the total number of materials, the number of material alarms, the number of material types, the number of material inventory alarms, and grasps the overall overview of the materials.

[0045] A computer-readable storage medium includes a stored program, wherein when the program is run, the device where the computer-readable storage medium is located is controlled to execute the distributed warehouse multi-source data centralized management system based on artificial intelligence as described above.

[0046] A processor is used to run a program, wherein when the program is run, it executes the distributed warehouse multi-source data centralized management system based on artificial intelligence as described above.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] 1. This invention addresses the needs of two types of orders by adopting two cargo storage modes to better handle both types of orders. A central dispatching unit allocates goods, either factory-produced or sourced from other sources, to distributed warehouses, recording all process and result data. Separate storage facilitates the storage and retrieval of goods for different orders and facilitates the acquisition of two different types of data for management.

[0049] 2. The present invention sets up a selection function for the specific storage warehouse of the goods, fully considering the distance between the customer address and the distributed warehouse unit address and the remaining storage capacity of the reserved storage area in the distributed warehouse unit. By solving the optimal storage warehouse for the reserved order goods, while fully considering the warehouse storage capacity, it also minimizes the customer's pickup distance as much as possible, facilitating the management of the warehouse's own inventory and pickup capacity, and maximizing customer satisfaction as much as possible.

[0050] 3. Regarding the specific storage locations of goods, the present invention considers the pick-up time and capacity of the user's scheduled order when determining the specific location of goods in the reserved storage area. Goods with earlier pick-up times should be placed earlier than those with later pick-up times. When determining the specific location of goods in the basic storage area, the popularity of the goods type should be considered, with more popular goods being placed earlier. This approach further maximizes customer satisfaction and minimizes the time it takes for more customers to retrieve their goods from the warehouse.

[0051] 4. The present invention also trains a neural network model to predict future data. Based on historical data trends, it predicts the time and quantity of future order pick-up, thereby better planning the storage space for each category in each predetermined storage unit. For example, if a large category has a large number of orders, the storage capacity for that category will need to be appropriately expanded. This approach further refines the specific classification of the storage units. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0053] Figure 1 This is a schematic diagram of the overall module of Example 1 of the present invention;

[0054] Figure 2 Schematic diagram of the structural framework of the general dispatching unit of the present invention;

[0055] Figure 3 This is a schematic diagram of the material management unit framework in Example 4 of the present invention. DETAILED DESCRIPTION

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] It will be understood that when used in this specification and the appended claims, the terms “comprises” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0058] It should also be understood that the terms used in the present specification are only for the purpose of describing particular embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0059] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0060] Example 1

[0061] In one embodiment of the present invention, a distributed warehouse multi-source data centralized management system based on artificial intelligence is provided, comprising at least one general scheduling unit and at least two distributed warehouse units and an execution unit, wherein:

[0062] The general dispatching unit is used to obtain user orders and the storage status of each distributed warehouse unit, and instruct the factory source to distribute and deliver goods to each distributed warehouse, or allocate goods between each distributed warehouse unit. In addition, the general dispatching unit also gives the specific storage location of the goods based on the different pick-up times of the user's scheduled orders; the general dispatching unit is provided with a management port for connecting to the terminal equipment of the management level to transmit data and receive instructions.

[0063] A distributed warehouse unit is used to store goods. The distributed warehouse unit is divided into a predetermined storage area and a basic storage area. The predetermined storage area is used to store goods required based on user predetermined orders, and the basic storage area is used to store goods directly allocated by the general scheduling unit.

[0064] The execution unit is used to receive instructions from the general scheduling unit and transport the goods to the specific storage location in the distributed warehouse unit.

[0065] The three units are described in further detail below.

[0066] In actual goods supply and demand transactions, demanders typically place orders in two formats: real-time orders, which typically require immediate pickup or the sooner the better. Pre-orders, which typically come with a specific pickup date, such as "pick up on XX / XX / XX," "pick up on the first business day of each month," or "pick up on the first Saturday of the fourth quarter." These orders all come with a specific future date, rather than real-time pickup or immediate pickup. Therefore, adopting two storage models can better address these two types of orders.

[0067] Therefore, in one embodiment of the present invention, a general scheduling unit is set up to distribute goods, distribute goods produced by factories or from other sources of goods to various distributed warehouses, and record all process data and result data in the process. The process data includes user orders (including user real-time orders and user scheduled orders) and goods allocation plans (including which specific goods from where are distributed to which distributed warehouses), and the result data is the outbound goods data, inbound goods data and goods data currently stored in the warehouse of each distributed warehouse.

[0068] In a distributed warehouse, two types of storage areas are used for goods: a basic storage area and a scheduled storage area. The basic storage area is used to store goods delivered directly from the factory or other source of goods, and is intended for immediate access by customers placing real-time orders. The transportation of goods in this area is subject to a certain degree of real-time and randomness. The scheduled storage area is used to store goods required based on user pre-orders, specifically to fulfill pre-orders. Each item in this area has a specific pickup time and pre-order information, including the pickup customer. The goods in this area are deterministic, with each item having a fixed delivery destination and delivery time. Separating storage into two areas facilitates the management of different orders and the management of the distributed warehouse. Separating storage into two areas allows the management end (the master dispatch unit) to monitor the operating conditions of different user orders in real time based on the order and storage conditions of the two areas. This data, such as the backlog rate, order volume, shipment volume, allocation volume, allocation costs, and storage costs of the two areas, facilitates management decisions. In short, separate storage not only facilitates the storage and retrieval of goods for different orders, but also facilitates the acquisition of two different data for management personnel to manage.

[0069] The general dispatching unit is mainly used to distribute the goods to different distributed warehouse units through the execution unit. The execution unit includes transport vehicles, transport personnel and other automated equipment, among which the automated equipment includes unmanned forklifts and dispatching systems (such as RCS), automated shelf systems and PLC control, omnidirectional pallet robots and other automated equipment that can be controlled by programs. In short, the role of the execution unit is only to transport the goods to the designated location according to the deployment plan of the general dispatching unit. It should be understood that in this process, the means of transportation may include manual, semi-manual and fully automatic. Based on different basic equipment and actual conditions, different methods are used to transport the goods to the designated location. The present invention does not make specific limitations here.

[0070] Example 2

[0071] In one embodiment of the present invention, based on Example 1, the cargo allocation process of the general scheduling unit is further provided:

[0072] The overall dispatching unit includes a data interaction module, a data storage module, a warehouse matching module, a location matching module, an alarm module and a decision-making module.

[0073] The data interaction module is used to obtain real-time customer orders and scheduled customer orders, and output the final transportation plan to the execution unit, playing the role of data interaction and transmission. Among them, the information in the real-time customer order includes at least customer information and order information, the customer information includes at least the customer address, and the order information includes at least the type and quantity of goods. In addition to customer information and order information, the customer scheduled order also includes the expected pickup time. The transportation plan output to the execution unit includes the cargo pick-up location and the cargo delivery location. The cargo pick-up location includes the factory and other cargo sources, as well as distributed warehouse units. The cargo delivery location is each distributed warehouse unit.

[0074] The data storage module stores the address of each distributed warehouse unit.

[0075] The warehouse matching module selects the specific distributed warehouse unit for storing the goods corresponding to the customer's scheduled order based on the address of the customer's scheduled order. The specific selection function is as follows:

[0076]

[0077] Where Z is the matching function, S 间距 is the distance between the customer address and the distributed warehouse unit address, R 预定区_剩余 The remaining storage capacity of the reserved storage area in the distributed warehouse unit.

[0078] For the matching function, there are residual capacity constraints and spacing constraints to avoid the problem that the residual capacity of distributed warehouse units is insufficient to accommodate goods and the spacing is too large.

[0079] The remaining capacity constraints are as follows:

[0080] R 剩余 ≥R 货物 +δ

[0081] Where R 货物 is the capacity of the goods, δ is the redundancy coefficient, which is set as a constant with a value range of δ ≥ 0. The setting of the redundancy coefficient is mainly to allow the distributed warehouse unit to reserve a certain amount of storage space to store goods ordered by other users.

[0082] The spacing constraints are as follows:

[0083] S 间距 ≤S 间距_max

[0084] Where S 间距_max is the maximum tolerance distance.

[0085] In summary, when selecting a specific distributed warehouse, the specific information of each warehouse is substituted into the matching function and solved to obtain the optimal distributed warehouse unit for storage. During this process, if there is no solution, the redundancy coefficient δ is further reduced and the solution is repeated. If the redundancy coefficient δ is reduced to 0 and there is still no solution, the alarm module is activated and an alarm message is sent to the administrator terminal so that the relevant staff can further process the corresponding customer's pre-order.

[0086] After determining the specific distributed warehouse where the goods are stored, it is necessary to further determine the specific storage location of the goods in the warehouse. In the general scheduling unit, the location matching module is used to select the specific location in the distributed warehouse. In the distributed warehouse unit, the predetermined storage area is divided into several storage departments. The storage departments are classified according to different storage capacities (such as large, medium, and small). Each storage department under each category has a unique number, and there are differences in the order between the storage departments under each category. The difference in order can be reflected in the unique number. For example, the number sequence of each storage department in the large category is X = [x1, x2,…, x i ,…,x n ], where n is the maximum n storage units of the predetermined storage area or basic storage area of the distributed warehouse unit, and i is the storage unit with the order i. The earlier the number, the easier it is to access. Easier access includes being closest to the warehouse door or having a relatively low height for easy access. The specific coding is based on the location of each warehouse and is not limited here. The numbers of other categories are the same as above.

[0087] Since the distributed warehouse unit is divided into two areas, and these two areas correspond to different customer orders and different data calculation methods, these two areas can be regarded as relatively independent areas under a warehouse, and the subsequent determination of the specific location of the goods in these two areas is also relatively independent.

[0088] When determining the specific location of goods in a reserved storage area, the pickup time and cargo capacity of the user's scheduled order must be taken into consideration. Goods with earlier pickup times should be placed before those with later pickup times. Therefore, determining the location of goods in a reserved storage area includes the following steps (the locations mentioned below are all reserved storage area locations and are not related to the basic storage area):

[0089] S1: Determine the storage department category for the goods based on their capacity. The storage categories include large, medium or small categories (or other more categories). After determining the storage department category, such as large category, select a storage department from several storage departments under this category to store the goods.

[0090] S2: Get the pick-up time of the scheduled order to be assigned, and determine whether there are goods stored in the storage unit of the current category. If so, get the specific pick-up time and placement location of the goods with the smallest difference from the pick-up time of the scheduled order to be assigned as the reference time and reference location, and then proceed to step S3; if not, store the goods of the scheduled order in the storage unit numbered x. 2 / n (If n is an odd number, 2 / n is an integer), that is, it is stored in the middle position of the sequence;

[0091] S3: Determine whether the pickup time of the scheduled order to be allocated is before or after the reference time. If it is before the reference time, the goods position of the scheduled order to be allocated that is in the first N positions of the reference position sequence is used as the position to be allocated. Otherwise, the goods position of the scheduled order to be allocated that is in the last N positions of the reference position sequence is used as the position to be allocated. The specific value of N is as follows:

[0092]

[0093] Where, T 待分配 is the pickup time of the order to be assigned, T 参考 is the reference time, T 间隔 is the average interval time of the storage department.

[0094] Average storage interval time T 间隔 Determined by:

[0095]

[0096] Where, T maxis the pickup time with the largest difference from the current time in the current storage department (the longest pickup time), T min N is the minimum pickup time (fastest pickup) in the current storage department. max-min T max and T min The interval between the specific storage locations of the goods of the corresponding order, that is, how many storage units are left empty in between.

[0097] S4: Determine the specific location of the goods to be allocated based on N calculated in step S3. If there is a decimal point after the time interval N calculated in step S3, round it up to the nearest integer. If the location where the goods to be allocated are already stored, postpone it to an empty space in the storage department. If there is no space left, postpone it to an earlier location.

[0098] When determining the specific location of goods in the basic storage area, the popularity of the goods type should be considered, and more popular goods should be placed in the front. The goods in the basic storage area are relatively stable. Since they are for real-time orders or offline purchase and pickup directly at the warehouse, the goods stored in the basic storage area are based on the quantity pre-arranged by management. Therefore, there is no fixed-capacity storage unit under the basic storage area. Instead of storing goods based on pickup time and order capacity, they are stored directly by popularity (that is, the number of people picking up the goods). Frequently purchased and picked-up goods are placed in the front of the storage location and sorted by historical order volume from high to low, corresponding to the storage location from front to back.

[0099] Example 3

[0100] This embodiment further improves upon Example 2 by further processing the data in Example 2 using a deep learning model, further refining the specific classification of the storage units. The general approach is to predict future order pickup times and quantities based on historical data trends, thereby better planning the storage space for each category within each predetermined storage unit. For example, if a large category has a large number of orders, the storage capacity for that category may need to be appropriately expanded.

[0101] Next, we will train a deep learning model to predict the pickup time and quantity of future orders. The specific steps are as follows:

[0102] S01: Collect historical order data, including order number, pickup time, pickup quantity, order time, customer information, product type, season, holidays, and other factors that may affect pickup time and quantity; remove duplicate data, handle missing values (such as filling or deleting), and eliminate outliers (such as detecting outliers through box plots); extract time features (such as week, month, quarter), customer features (such as customer type, historical order quantity), and product features (such as product category, price);

[0103] S02: Divide the dataset into training and test sets, usually in chronological order, such as using the first 80% of the data as the training set and the last 20% as the test set; normalize or standardize numerical features to facilitate better learning of the model;

[0104] S03: Convert time series data into a format suitable for LSTM input. For example, organize the features of each time step (such as pickup quantity, time features, etc.) into a sequence and determine the time steps. For example, use data from the past five years to predict the pickup time and quantity for the next year.

[0105] S04: Normalize or standardize numerical features (such as pickup quantity, price, etc.) so that the data is distributed in the range of [0, 1] or with a mean of 0 and a standard deviation of 1. Record the normalized parameters (such as maximum and minimum values or mean and standard deviation) so that the input data and output results can be denormalized during prediction.

[0106] S05: Build an LSTM model, as follows:

[0107] Input layer: defines the input shape, for example (timesteps, features), where timesteps is the time step and features is the number of features per time step;

[0108] LSTM layer: Add an LSTM layer, set the appropriate number of units (such as 64, 128, etc.), and choose whether to return sequences (return_sequences = True or False); multiple LSTM layers can be stacked to enhance the expressiveness of the model.

[0109] Fully connected layer: Add a fully connected layer (Dense layer) after the LSTM layer to map the LSTM output to the prediction target. For delivery time prediction, the activation function of the output layer can be set to linear; for delivery quantity prediction, a linear activation function can also be used.

[0110] Compile the model: Select an optimizer (such as Adam or RMSprop). Define the loss function. For regression tasks, the mean squared error (MSE) or mean absolute error (MAE) is usually used. Set the evaluation metric, such as MAE or RMSE.

[0111] S06: Convert the training set data into the input format required by LSTM, such as (samples, timesteps, features). Convert the target variables (pickup time and pickup quantity) into a format suitable for model training. Set appropriate hyperparameters, such as learning rate, batch size, and number of epochs. Train the model using the training set data and evaluate performance on the validation set. Use early stopping to prevent overfitting, for example, stopping training when the validation set loss does not improve within 10 consecutive epochs.

[0112] S07: Use indicators such as root mean square error (RMSE), mean absolute error (MAE), and symmetric mean absolute percentage error (SMAPE) to evaluate the prediction performance of the model.

[0113] S08: Use GridSearchCV or RandomSearchCV to adjust LSTM hyperparameters, such as the number of units, learning rate, and batch size. Try different timesteps and feature combinations to find the optimal model configuration. Add dropout layers to LSTM or fully connected layers to reduce the risk of overfitting. Use L2 regularization to constrain the model weights.

[0114] S09: Use the selected model to make predictions and transmit the predicted data to the management terminal.

[0115] In summary, the present invention addresses the needs of two types of orders by adopting two cargo storage modes, enabling targeted processing of both types of orders. A central dispatching unit is deployed to allocate goods, assigning factory-produced goods or goods from other sources to various distributed warehouses and recording all process and result data. Separate storage facilitates the storage and retrieval of goods for different orders while also facilitating the acquisition of two distinct data types for management. Furthermore, a selection function is implemented for the specific storage warehouses for goods, taking into account the distance between the customer's address and the address of the distributed warehouse unit, as well as the remaining storage capacity of the designated storage areas within the distributed warehouse units. By solving this problem, the optimal storage warehouse for goods for a predetermined order is determined, minimizing the distance customers need to reach while taking into account warehouse storage capacity. This facilitates inventory and retrieval capacity management within the warehouse, while maximizing customer satisfaction. Regarding the specific storage locations of goods, the designated storage areas should consider the pickup time and cargo capacity of the customer's scheduled order. Goods with earlier pickup times should be placed in front of those with later pickup times. When determining the specific location of goods in the basic storage area, the popularity of the goods should be taken into consideration, and the more popular goods should be placed at the front. In this way, customer satisfaction is further maximized and as many customers as possible can reduce the time it takes to retrieve goods from the warehouse.

[0116] Example 4

[0117] This embodiment, based on any one of Embodiments 1-3, further includes a material management unit, which includes two components: a terminal and a material management system.

[0118] The terminal is used to display incoming and outgoing warehouse information and inventory records, and to compile statistics on the number of material types in the warehouse, the number of material inventory warnings, the number of new materials added this year, the number of incoming material orders this year, and the number of outgoing material orders this year to understand the warehouse situation. In addition, facial recognition and facial recognition training are also performed.

[0119] The material management system displays the distribution of all materials under the jurisdiction of the unit based on the Zhikan map, counts the total number of materials, the number of material alarms, the number of material types, the number of material inventory alarms, and grasps the overall overview of the materials.

[0120] The details are shown in the following table:

[0121]

[0122]

[0123]

[0124]

[0125] Those skilled in the art will appreciate that the units of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0126] In the embodiments provided by the present invention, it should be understood that the division of units is merely a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0127] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0128] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nly Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc., various media that can store program code.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. A distributed warehouse multi-source data centralized management system based on artificial intelligence, characterized by: It includes at least one general scheduling unit and at least two distributed warehouse units and execution units, wherein: The general dispatching unit is used to obtain user orders and the storage status of each distributed warehouse unit, and instruct the factory source to distribute and deliver goods to each distributed warehouse, or allocate goods between distributed warehouse units. In addition, the general dispatching unit also provides the specific storage location of the goods based on the different pick-up times of user scheduled orders; The distributed warehouse unit is used to store goods. The distributed warehouse unit is divided into a predetermined storage area and a basic storage area. The predetermined storage area is used to store goods required for user predetermined orders in a sequential order based on the pick-up time and the ordered goods. The basic storage area is used to store goods directly allocated by the general scheduling unit. The execution unit is used to receive instructions from the general scheduling unit and transport the goods to the specific storage location in the distributed warehouse unit.

2. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to claim 1, characterized in that: The general dispatching unit distributes goods, distributes goods produced by factories or from other sources to various distributed warehouses, and records all process data and result data in the process. The process data includes user orders and goods allocation plans, and the result data is the outbound goods data, inbound goods data and goods data currently stored in each distributed warehouse.

3. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to claim 2 is characterized in that: In a distributed warehouse, there are two types of cargo storage areas, one is the basic storage area and the other is the reserved storage area. The basic storage area is used to store cargo that is directly delivered from the factory or other cargo sources, and is directly available to customers with real-time orders. The reserved storage area is used to store cargo required based on user reserved orders. Each cargo has a fixed output object and output time.

4. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to claim 1, characterized in that: The general dispatching unit includes data interaction module, data storage module, warehouse matching module and alarm module; The data interaction module is used to obtain real-time customer orders and pre-ordered customer orders, and output the final transportation plan to the execution unit; wherein the information in the real-time customer orders includes at least customer information and order information, the customer information includes at least the customer address, and the order information includes at least the type and quantity of goods; in addition to the customer information and order information, the pre-ordered customer orders also include the expected pickup time; The data storage module stores the address of each distributed warehouse unit; The warehouse matching module selects a specific distributed warehouse unit for storing the goods corresponding to the customer's scheduled order based on the address of the customer's scheduled order; The alarm module is used to send an alarm message to the manager terminal when the warehouse cannot select a specific distributed warehouse unit.

5. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to claim 4 is characterized in that: The warehouse matching module has a selection function. By solving the selection function, the specific storage warehouse of the goods is obtained. The specific selection function is as follows: Where Z is the matching function, S 间距 is the distance between the customer address and the distributed warehouse unit address, R 预定区_剩余 The remaining storage capacity of the reserved storage area in the distributed warehouse unit; For the matching function, there are remaining capacity constraints and spacing constraints; The remaining capacity constraints are as follows: R 剩余 ≥R 货物 Where R 货物 The capacity of the cargo; The spacing constraints are as follows: S 间距 ≤S 间距_max Where S 间距_max is the maximum tolerable spacing.

6. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to claim 5, characterized in that: The redundancy factor is considered in the remaining capacity constraint, as follows: R 剩余 ≥R 货物 +d Where R 货物 is the capacity of the cargo, δ is the redundancy coefficient, which is set as a constant and has a value range of δ≥0; When selecting a specific distributed warehouse, the specific information of each warehouse is substituted into the matching function for solution to obtain the optimal distributed warehouse unit for storage; in this process, if there is no solution, the value of the redundancy coefficient δ is further reduced and the solution is performed again. If the redundancy coefficient δ has dropped to 0 and there is still no solution, the alarm module is activated and an alarm message is sent to the administrator terminal.

7. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to claim 4 is characterized in that: The general scheduling unit also includes a location matching module, which is used to select a specific location in the distributed warehouse. In the distributed warehouse unit, a predetermined storage area is divided into several storage units. The storage units are classified according to different storage capacities. Each storage unit in each category has a unique number, and there is a difference in the order of the storage units in each category. The specific storage location of the reserved storage area is determined as follows: S1: Determine the storage category of the goods according to their capacity. Once the storage category is determined, select one storage category from among several storage categories under the category to store the goods. S2: Get the pick-up time of the scheduled order to be assigned, and determine whether there are goods stored in the storage unit of the current category. If so, get the specific pick-up time and placement location of the goods with the smallest difference from the pick-up time of the scheduled order to be assigned as the reference time and reference location, and then proceed to step S3; if not, store the goods of the scheduled order in the storage unit numbered x. 2 / n , that is, store it in the middle of the sequence; S3: Determine whether the pickup time of the scheduled order to be allocated is before or after the reference time. If it is before the reference time, the goods position of the scheduled order to be allocated that is in the first N positions of the reference position sequence is used as the position to be allocated. Otherwise, the goods position of the scheduled order to be allocated that is in the last N positions of the reference position sequence is used as the position to be allocated. The specific value of N is as follows: Where, T 待分配 is the pickup time of the order to be assigned, T 参考 is the reference time, T 间隔 is the average interval time of the storage department; Average storage interval time T 间隔 Determined by: Where, T max is the delivery time with the largest difference from the current time in the current storage department, T min N is the minimum delivery time in the current storage department and the current time. max-min T max and T min The interval between the specific storage locations of the goods of the corresponding order, that is, how many storage units are left empty in between.

8. The distributed warehouse multi-source data centralized management system based on artificial intelligence according to any one of claims 1 to 7, characterized in that: It also includes a material management unit, which includes two components: a terminal and a material management system, wherein: The terminal is used to display incoming and outgoing warehouse information and inventory records, and to compile statistics on the number of material types in the warehouse, the number of material inventory warnings, the number of new materials added this year, the number of incoming material orders this year, and the number of outgoing material orders this year to understand the warehouse situation. In addition, facial recognition and facial recognition training are also performed. The material management system displays the distribution of all materials under the jurisdiction of the unit based on the Zhikan map, counts the total number of materials, the number of material alarms, the number of material types, the number of material inventory alarms, and grasps the overall overview of the materials.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the distributed warehouse multi-source data centralized management system based on artificial intelligence as described in any one of claims 1 to 8.

10. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the distributed warehouse multi-source data centralized management system based on artificial intelligence as described in any one of claims 1 to 8.

Images