A big data-based warehouse logistics intelligent management method and system

Abstract

The present application relates to the technical field of warehouse logistics intelligent management, in particular to a warehouse logistics intelligent management method and system based on big data, comprising: collecting full-link original operation data, generating structured feature data set containing cargo life cycle, equipment performance matrix and order space-time distribution after multi-dimensional fusion processing; based on the pre-constructed warehouse logistics simulation model, the data set is strategy deduced, the dynamic management strategy set is generated and converted into executable control signal; after the system is executed, the operation feedback data is monitored in real time, and the simulation model parameters are iteratively updated based on the operation feedback data, forming a management closed loop. Through the simulation deduction and model self-updating mechanism, the method realizes the forward-looking optimization of warehouse strategy and the continuous self-adaptation of the system, and improves the overall operation efficiency and decision accuracy of the warehouse logistics system in complex dynamic environment.

Description

Technical Field

[0001] This invention relates to the field of intelligent management technology for warehousing and logistics, and in particular to an intelligent management method and system for warehousing and logistics based on big data. Background Technology

[0002] Existing warehouse logistics management primarily relies on automated systems based on fixed rules or real-time data-driven scheduling algorithms. These technologies trigger preset response rules or perform localized, real-time optimization calculations by monitoring inventory levels, equipment status, and order queues in real time. However, warehouse operations are complex systems with multiple constraints, multiple objectives, and a dynamically changing environment. Static rules cannot adapt to real-time fluctuations in order patterns and equipment status, while scheduling optimizations focused solely on the current moment lack the ability to simulate and evaluate the medium- to long-term effects of different strategies. This results in management strategies often being limited to local or single-point optimizations, with strategies between different operational stages potentially lacking coordination or even conflicting at the global level, making it difficult to achieve consistently optimal overall system efficiency.

[0003] The models or rule parameters supporting management decisions are typically set during system deployment, with subsequent updates relying on human experience and periodic offline analysis. Equipment performance, cargo storage characteristics, and demand patterns in a warehousing system all change slowly or drastically over time. Fixed decision-making models cannot perceive or adapt to this continuous environmental drift, and their decision accuracy naturally declines with operating time. System performance thus suffers from a degradation that is difficult to automatically repair until manual intervention is needed to reset the model or perform a major parameter update. This disconnect between the model and the environment creates a bottleneck in the system's ability to maintain efficient and accurate operation over the long term. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a big data-based intelligent management method and system for warehousing and logistics.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a big data-based intelligent management method for warehousing and logistics, comprising: Collect raw operational data from the entire warehousing and logistics system. This raw operational data includes real-time information on goods entering the warehouse, monitoring data on the storage status within the warehouse, equipment operation logs, and outbound order flow data. The original operational data is subjected to multi-dimensional cross-fusion processing to generate a structured warehousing feature dataset, which includes cargo storage lifecycle characteristics, warehousing equipment performance status matrix, and order demand spatiotemporal distribution map. Based on the pre-built warehousing and logistics simulation model, strategy deduction and analysis are performed on the structured warehousing feature dataset to generate a set of dynamic management strategies for the warehousing and logistics system. The set of dynamic management strategies includes inventory optimization configuration schemes, equipment scheduling path instruction sets, and work process reconstruction strategies. The set of dynamic management strategies is input into the warehouse execution control unit to generate an executable control signal queue, which includes a location allocation instruction sequence, a handling equipment control sequence, and an operation node status switching instruction. The system monitors the operational feedback data of the warehousing and logistics system after executing the control signal queue in real time, and uses the operational feedback data to iteratively update the parameters of the warehousing and logistics simulation model.

[0006] As a further aspect of the present invention, the step of performing multi-dimensional cross-fusion processing on the original operational data to generate a structured warehouse feature dataset includes: Extract cargo attribute tags and cargo entry time stamps from the real-time cargo entry information to construct a cargo file table with the cargo unique code as the primary key; The storage status monitoring data in the warehouse is associated and mapped with the cargo file table to generate a storage history trajectory chain for each cargo, which includes the current location, storage duration, and environmental parameter records. The equipment operation logs are analyzed to extract the location sequence, energy consumption data and fault warning events of the handling equipment, and to generate a performance status matrix of the warehousing equipment that represents the health and load status of the equipment. The order information in the outbound order flow data is aggregated, and cluster analysis is performed according to the predetermined time granularity and spatial region to draw the spatiotemporal distribution map of the order demand, which reflects the distribution of order volume, goods type and target flow direction. The storage history trajectory chain, the performance status matrix of the warehousing equipment, and the spatiotemporal distribution map of order demand are spatiotemporally aligned and associated with features to form the structured warehousing feature dataset.

[0007] As a further aspect of the present invention, the step of parsing the equipment operation log, extracting the location sequence, energy consumption data, and fault warning events of the handling equipment, and generating the performance status matrix of the warehousing equipment characterizing the health and load status of the equipment includes: Identify the movement command records and sensor feedback data of each handling device in the equipment operation log, and construct a device position sequence including timestamps, coordinates, and speeds; Extract energy consumption data of the handling equipment under different operating modes from the log power consumption records, and calculate its energy efficiency index per unit of work. The fault diagnosis rule engine is used to scan the device's operation logs and identify events that meet the characteristics of fault precursors as fault warning events. Based on the regularity of the equipment location sequence, the stability of the energy efficiency index, and the frequency of the fault warning events, a comprehensive health score for each handling device is generated using a weighted scoring model. The overall health score and current load information of all handling equipment are summarized to form a multi-row, multi-column performance status matrix of the warehousing equipment.

[0008] As a further aspect of the present invention, the pre-built warehousing and logistics simulation model is used to perform strategy deduction and analysis on the structured warehousing feature dataset to generate a set of dynamic management strategies for the warehousing and logistics system, including: The structured warehousing feature dataset is used as the initial state parameter and loaded into the warehousing and logistics simulation model; In the warehousing and logistics simulation model, multiple simulation scenarios are set up, and each simulation scenario corresponds to a different set of inventory strategies, equipment scheduling rules and work process assumptions. The warehousing and logistics simulation model is driven to run all simulation scenarios in parallel, simulating the evolution of the warehousing and logistics system's operating status under various strategy assumptions over a period of time in the future. During the simulation, key performance indicators (KPIs) for each simulation scenario are calculated and recorded in real time. These KPIs include warehouse capacity utilization, average order fulfillment time, and overall equipment utilization. By comparing and analyzing the key performance indicator data after the completion of all simulated scenarios, the inventory strategy, equipment scheduling rules and work process assumptions corresponding to the optimal values ​​are selected, and the inventory optimization configuration scheme, equipment scheduling path instruction set and work process reconstruction strategy are integrated to generate the aforementioned inventory optimization configuration scheme, equipment scheduling path instruction set and work process reconstruction strategy.

[0009] As a further aspect of the present invention, the step of inputting the dynamic management strategy set into the warehouse execution control unit to generate an executable control signal queue includes: The inventory optimization configuration scheme is analyzed, and the goods storage suggestions are converted into specific inbound, transfer or inventory count instructions. The target storage location for each instruction is determined based on the current storage location occupancy, forming a storage location allocation instruction sequence. The equipment scheduling path instruction set is matched with the warehouse electronic map and the real-time location of the equipment. Execution equipment is assigned to each scheduling path, and a conflict-free movement path time window is planned to generate the control sequence of the handling equipment. The workflow reconstruction strategy is broken down into the state transition conditions and operation steps of each work node. Based on the current system work progress, the corresponding node state change is triggered to form a work node state switching instruction. According to the preset execution priority and timing dependency, the sequence of storage location allocation instructions, the timing sequence of handling equipment control, and the state switching instructions of operation nodes are sorted and encapsulated to generate the executable control signal queue.

[0010] As a further aspect of the present invention, the inventory optimization configuration scheme is analyzed, and the goods storage suggestions therein are converted into specific inbound, transfer, or inventory counting instructions. The target storage location for each instruction is determined based on the current storage location occupancy, forming a storage location allocation instruction sequence, including: Identify the recommendations regarding cargo storage density and storage location optimization in the aforementioned inventory optimization configuration scheme; For goods recommended for warehousing, the required storage space is calculated based on the goods attribute tags, and candidate storage locations that meet the storage conditions are selected from the set of available storage locations. The target storage location is selected according to the preset storage location selection rules, and an warehousing instruction is generated. For goods that are recommended to be moved, based on the access frequency in their storage history trajectory chain, they are reassigned from their current inefficient storage location to an efficient storage location, and a move instruction and the new and old storage location addresses are generated. For goods that need to be inventoried, an inventory instruction is generated and the priority of the inventory is specified based on their storage time or special storage requirements. The generated inbound instructions, transfer instructions, and inventory instructions are arranged according to the urgency of the business and the order of execution logic to form the storage location allocation instruction sequence.

[0011] As a further aspect of the present invention, the real-time monitoring of the operational feedback data of the warehousing and logistics system after executing the control signal queue, and the use of the operational feedback data to iteratively update the parameters of the warehousing and logistics simulation model, includes: Deploy a sensor network and data acquisition interface to capture system response data generated during the execution of the control signal queue in real time. The system response data includes the instruction completion status, the actual operating trajectory and completion time of the equipment, and the actual occupancy change of the cargo space. The system response data is compared with the expected execution state in the control signal queue to calculate the deviation metric of the strategy execution. The deviation metric reflects the difference between the preset effect and the actual effect of the strategy. Extract new features from the system response data that reflect dynamic changes in the warehousing environment and equipment performance drift to form a model feedback feature set; Using the deviation metric and the model feedback feature set as training samples, the parameters of the strategy deduction part in the warehousing and logistics simulation model are incrementally adjusted to reduce the difference between the model deduction results and the actual results. After the parameters are adjusted, the updated warehousing and logistics simulation model will be used in the next round of management strategy generation.

[0012] As a further aspect of the present invention, the step of comparing the system response data with the expected execution state in the control signal queue and calculating the deviation metric of strategy execution includes: Extract the expected start and end points and expected completion time of each instruction from the control signal queue; Match the actual start and end points of each instruction with the actual completion time from the system response data; Calculate the spatial offset between the actual execution start point and the expected execution start point, and between the actual end point and the expected end point for each instruction; Calculate the time deviation between the actual completion time and the expected completion time for each instruction; The spatial offsets and temporal deviations of all instructions are aggregated, and a weighted average method is used to calculate a comprehensive deviation metric.

[0013] As a further aspect of the present invention, the method also includes a dynamic strategy triggering mechanism based on order demand: Continuously monitor the incoming outbound order flow data and analyze the characteristic patterns of new orders in real time; When the type, quantity, or urgency of goods in a new order constitutes a predetermined characteristic pattern, the strategy regeneration process is triggered. In the strategy regeneration process, the new order data is integrated into the structured warehouse feature dataset to form an updated dataset; Based on the updated dataset, the strategy deduction and analysis of the warehousing and logistics simulation model are re-executed to generate a set of dynamic management strategies for the arrival of new orders. The newly generated set of dynamic management strategies is transformed into an executable control signal queue and sent to the warehouse execution control unit for priority execution.

[0014] As a further aspect of the present invention, the present invention also includes a big data-based intelligent warehousing and logistics management system. The system includes a data processing server and a warehouse control gateway. The data processing server and the warehouse control gateway exchange data. The data processing server is used to store program instructions and model parameters. When the data processing server executes the program instructions, it implements all the steps of the above-described big data-based intelligent warehousing and logistics management method.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Based on a pre-built warehousing and logistics simulation model, this method utilizes a structured dataset incorporating multi-dimensional features for strategy extrapolation and analysis. This enables systematic simulation and evaluation of strategy combinations, including inventory configuration, equipment scheduling, and operational processes, within a virtual environment. This approach allows the system to predict and compare the interactive effects and long-term operational outcomes of various potential management strategies before actual physical execution. The dynamically generated management strategy set comprehensively considers multiple constraints, such as the lifecycle of goods storage, real-time equipment performance, and the spatiotemporal distribution of orders, thus surpassing decision-making models that rely on single rules or real-time calculations. The resulting optimized inventory configuration schemes, equipment scheduling path instruction sets, and operational process reconfiguration strategies possess better global coordination and forward-looking adaptability, enabling them to cope with complex and fluctuating warehousing operation environments.

[0016] By collecting real-time operational feedback data after the system executes control signal queues, and using this data to iteratively update the parameters of the warehousing and logistics simulation model, a self-evolving closed-loop optimization system is constructed. The operational feedback data continuously reflects dynamic information such as actual equipment performance degradation, changes in operational efficiency, and real order flow patterns. The automatic iteration of model parameters ensures that the internal logic of the simulation model evolves synchronously with the operational characteristics of the actual physical system. This overcomes the problem of fixed models gradually becoming ineffective due to environmental drift, ensuring that the model basis for strategy derivation remains consistent with the real-world state. Therefore, the system's management strategy can continuously self-calibrate and optimize over time, achieving the goal of maintaining decision-making accuracy and system operational efficiency in a dynamic environment. Attached Figure Description

[0017] Figure 1 The flowchart shows the intelligent warehousing and logistics management method based on big data as described in this invention. Figure 2 A flowchart for generating a structured warehouse feature dataset; Figure 3 A flowchart for generating a dynamic management strategy set for strategy deduction and analysis; Figure 4 Trend monitoring chart for core performance indicators of the intelligent management system for warehousing and logistics; Figure 5 This is a comparison chart of dynamic monitoring of multi-rule matching degree and trigger threshold based on a 5-minute time window. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1 The system collects raw operational data from the entire warehousing and logistics system, including real-time information on goods entering the warehouse, in-warehouse storage status monitoring data, equipment operation logs, and outbound order flow data. This raw operational data undergoes multi-dimensional cross-fusion processing to generate a structured warehousing feature dataset. This dataset includes goods storage lifecycle characteristics, a warehousing equipment performance status matrix, and a spatiotemporal distribution map of order demand. Based on a pre-built warehousing and logistics simulation model, the system performs strategy deduction and analysis on the structured warehousing feature dataset, generating a set of dynamic management strategies for the warehousing and logistics system. This set of dynamic management strategies includes inventory optimization configuration schemes, equipment scheduling path instruction sets, and work process refactoring strategies. The dynamic management strategy set is input into the warehousing execution control unit to generate an executable control signal queue. This executable control signal queue includes a sequence of location allocation instructions, a timing sequence for handling equipment control, and work node status switching instructions. The system monitors the operational feedback data of the warehousing and logistics system after executing the control signal queue in real time and uses this feedback data to iteratively update the parameters of the warehousing and logistics simulation model.

[0021] See Figure 2Taking a specific scenario of an e-commerce warehousing center as an example, the warehousing and logistics system continuously generates raw operational data across the entire chain. Real-time information on goods entering the warehouse includes fields such as unique goods code, goods type, entry timestamp, and batch number. In-warehouse storage status monitoring data includes fields such as current shelf coordinates, temperature and humidity sensor readings, and inventory quantity change events. Equipment operation logs include fields such as handling equipment identification, motor current records, travel coordinate sequence, and fault codes. Outbound order flow data includes fields such as order number, order time, goods list, and delivery address. Goods attribute tags and entry timestamps are extracted from the real-time information on goods entering the warehouse. Goods attribute tags include, for example, goods size, weight, and storage temperature requirements. The entry timestamp is accurate to the millisecond level. A goods file table is constructed with the unique goods code as the primary key. Each record in the goods file table is associated with the basic attributes of the goods and the entry time. In some embodiments, the association mapping between in-warehouse storage status monitoring data and cargo file tables is achieved using the cargo's unique code as the association key. Each cargo location change or environmental monitoring event is recorded, generating a storage history trajectory chain for each cargo that includes the current three-dimensional coordinates, cumulative storage duration, and historical temperature and humidity sequences. Equipment operation logs are parsed to extract the location sequences, energy consumption data, and fault warning events of the handling equipment, generating a warehouse equipment performance status matrix characterizing equipment health and load status. Order information from outbound order flow data is aggregated, and cluster analysis is performed according to an hourly time granularity and the spatial regions of warehouse partitions A, B, and C to draw a spatiotemporal distribution map of order demand reflecting the distribution of order volume, cargo type, and target flow direction. It is understandable that the storage history trajectory chain reflects the dynamic flow path of goods within the warehouse, the warehouse equipment performance status matrix reflects the real-time working capacity of the handling equipment, and the order demand spatiotemporal distribution map reveals the clustering pattern of orders. The storage history trajectory chain, the warehouse equipment performance status matrix, and the order demand spatiotemporal distribution map are spatiotemporally aligned and feature-associated. The alignment operation is based on a unified time reference coordinate system and warehouse geographical coordinate system, and the association operation is based on the logical connection between the goods code and the equipment identifier, thus forming a structured warehouse feature dataset.

[0022] In some embodiments, the movement command records and sensor feedback data of each handling device in the equipment operation log are identified. The movement command records include the coordinates of the starting point and the target point, and the sensor feedback data includes the real-time positioning coordinates and velocity vector. A device position sequence containing timestamps, coordinates, and velocities is constructed, and the device position sequence is arranged in ascending order of time. Energy consumption data of the handling devices under different operating modes are extracted from the log power consumption records. The operating modes include unloaded travel, loaded travel, and lifting operations. The energy efficiency index per unit of work is calculated, and the unit of work is defined as the product of the handling distance and the weight of the goods. The equipment operation log is scanned using a fault diagnosis rule engine. The fault diagnosis rule engine predefines a series of event patterns, such as the pattern of motor current continuously exceeding the threshold, and identifies events that meet the characteristics of fault precursors as fault warning events. Optionally, considering the regularity of the device position sequence, the stability of the energy efficiency index, and the frequency of fault warning events, a comprehensive health score for each handling device is generated through a weighted scoring model. The weighted scoring model adopts a linear weighting form, and the specific formula is expressed as follows:

[0023] in: This represents the overall health score of the handling equipment. , , Represents the pre-configured weighting coefficients and , The regularity score representing the equipment location sequence is obtained by calculating the reciprocal of the standard deviation of the location sequence and normalizing it. The stability score, representing the energy efficiency index, is obtained by calculating the reciprocal of the coefficient of variation of energy efficiency within a continuous time window and normalizing it. The frequency of fault warning events is obtained by counting the number of events per unit time and normalizing it to the [0,1] interval. The overall health score of all handling equipment and its current load task information are summarized, including task type and estimated time, forming a multi-row, multi-column warehouse equipment performance status matrix. Each row of the matrix corresponds to one handling equipment, and the columns include equipment identifier, overall health score, current load task description, and current location coordinates.

[0024] Optionally, in implementation, data comparison is reflected in the changes in data structure before and after processing. The equipment operation log entries in the original operating data are time-series text records. After parsing and calculation, the resulting warehouse equipment performance status matrix is ​​a two-dimensional table, where the comprehensive health score is a scalar value between 0 and 1, used to quantify and compare the health status of different equipment. It can be understood that the association mapping between the cargo file table and the storage history trajectory chain allows scattered monitoring point data to be linked into an ordered trajectory, and the order demand spatiotemporal distribution map aggregates discrete order flows into a heatmap with spatiotemporal dimensions. These structured features provide direct input for subsequent simulations. In implementation, the spatiotemporal alignment operation requires synchronizing the timestamps in the storage history trajectory chain, the timestamps in the warehouse equipment performance status matrix, and the time windows in the order demand spatiotemporal distribution map. The feature association operation requires spatial overlay analysis of cargo storage locations with equipment working areas and high-order-occurrence areas, thereby generating a unified structured warehouse feature dataset.

[0025] See Figure 3 In one embodiment of the present invention, the warehousing and logistics simulation model is a digital system based on agent modeling and discrete event-driven processing. A structured warehousing feature dataset is loaded into the simulation model as initial state parameters. These initial state parameters include the current location and attributes of all goods in the goods storage lifecycle characteristics, the health scores and real-time locations of all handling equipment in the warehousing equipment performance status matrix, and order prediction data for a future period from the order demand spatiotemporal distribution map. Multiple simulation scenarios are set up in the warehousing and logistics simulation model. Each scenario corresponds to a different set of inventory strategies, equipment scheduling rules, and operational process assumptions. Differences in inventory strategies include storing goods according to turnover rate or storing them centrally by category; differences in equipment scheduling rules include prioritizing nearest-distance order dispatch or balanced load order dispatch; and differences in operational process assumptions include parallel order picking or wave-aggregated picking. The warehousing and logistics simulation model runs all simulation scenarios in parallel, simulating the evolution of the warehousing and logistics system's operational state under various strategy assumptions over a future period. The simulation process is advanced by a virtual clock, calculating the logic and time consumption of all events such as goods movement, equipment operation, and order processing in the virtual environment. During the simulation, key performance indicators (KPIs) for each scenario are calculated and recorded in real time. These KPIs include warehouse capacity utilization, average order fulfillment time, and overall equipment utilization. Warehouse capacity utilization is the ratio of occupied storage space to total storage space. Average order fulfillment time is the average virtual time from order creation to goods dispatch. Overall equipment utilization is the ratio of effective equipment operating time to the total simulation duration. After comparing and analyzing the KPI data from all simulation scenarios, the optimal inventory strategy, equipment scheduling rules, and workflow assumptions are selected. These are then integrated to generate an optimized inventory configuration scheme, an equipment scheduling path instruction set, and a workflow refactoring strategy.

[0026] In some embodiments, the comparative analysis process involves a comprehensive evaluation of multiple key performance indicators, specifically expressed by the following formula:

[0027] in: Representing the The overall performance score of each simulated scenario The storage capacity utilization rate represents this scenario. The average order fulfillment time represents this scenario. The overall utilization rate of equipment in this scenario , , These are pre-set positive weighting coefficients used to balance the importance of different indicators. This represents the reciprocal of the average order fulfillment time to convert it into a positive metric. This can be understood as comparing all simulated scenarios. The strategy combination corresponding to the simulation scenario with the highest value is determined as optimal. Optionally, the integrated inventory optimization configuration scheme is to convert the virtual storage layout and inventory level of goods at the end of the simulation in the optimal simulation scenario into specific storage suggestion text; the integrated equipment scheduling path instruction set is to convert the complete movement trajectory sequence of all handling equipment in the optimal simulation scenario within the simulation period into path point instructions with timestamps; and the integrated workflow reconstruction strategy is to convert the virtual workflow steps and node logic of order processing in the optimal simulation scenario into configurable process rules.

[0028] In some embodiments, the parallel operation of the warehousing and logistics simulation model relies on a high-performance computing cluster. Each simulation scenario runs on an independent computing node, and the number of simulation scenarios can be dozens or hundreds. In specific implementations, data comparison is reflected between input and output. The input is a unified structured warehousing feature dataset, and the output consists of multiple sets of key performance indicator data for different strategy assumptions and a final selected set of dynamic management strategies. By comparing the warehouse capacity utilization rate, average order fulfillment time, and overall equipment utilization rate under different simulation scenarios, the potential differences in the effectiveness of different management strategies can be quantitatively evaluated. It can be understood that using the structured warehousing feature dataset as the initial state parameter ensures that the starting point of the simulation is completely consistent with the state of the real warehouse at a certain moment. The simulation scenario settings cover multiple feasible strategy combinations, and parallel simulation improves the efficiency of strategy optimization. In specific implementations, the recording of key performance indicator data is periodic; for example, data is recorded after each virtual workday is simulated to observe the performance trend of the strategy during the simulation period.

[0029] In one embodiment of the present invention, the dynamic management strategy set exists in the form of a structured data file. The process of parsing the inventory optimization configuration scheme involves reading the strategy description field in the data file. The goods storage suggestions in the inventory optimization configuration scheme exist in the logical form of "goods code - suggested operation - target area", such as "goods code A123 - transfer - near sorting area B". Converting this into specific inbound, transfer or inventory instructions requires determining the target location for each instruction based on the current location occupancy. The current location occupancy is provided by the real-time inventory snapshot of the warehouse management system, forming a sequence of location allocation instructions. The equipment scheduling path instruction set is matched with the warehouse electronic map and the real-time location of the equipment. The warehouse electronic map is a digital map containing the coordinates of all shelves, aisles, and workstations. The real-time location of the equipment is obtained from the performance status matrix of the warehouse equipment. Assigning execution equipment to each scheduling path requires calculation based on the comprehensive health score of the handling equipment and its current location. Planning a conflict-free movement path time window requires calculating the expected arrival and departure times of each piece of equipment in each segment of the path, generating the handling equipment control sequence. The workflow refactoring strategy is broken down into state transition conditions and operational steps for each work node. The strategy defines new workflow steps such as "order merging - batch picking - review and packaging." Work node state switching instructions are control commands that trigger a node's state to change from "pending execution" to "in execution" or "completed." Based on preset execution priorities and timing dependencies (e.g., instructions related to urgent orders have higher priority), and timing dependencies (meaning certain instructions can only begin after other instructions are completed), the sequence of location allocation instructions, the timing of handling equipment control, and work node state switching instructions are sorted and encapsulated to generate an executable control signal queue.

[0030] In some embodiments, suggestions regarding goods storage density and storage location optimization in the inventory optimization configuration scheme are identified. Goods storage density suggestions include "compactly stacking similar goods," and storage location optimization suggestions include "storing frequently accessed goods in the exit area." For goods recommended for warehousing, the required storage space is calculated based on their goods attribute tags, which include length, width, height, and weight information. Candidate storage locations meeting the storage conditions are selected from the set of available storage locations. These conditions include storage location load-bearing capacity, size, and temperature zone. A target storage location is selected according to a preset storage location selection rule, which can be an evaluation function, generating an warehousing instruction. For goods recommended for relocation, considering the access frequency in their storage history trajectory chain (which records historical access timestamps), they are reassigned from their current inefficient storage location to a high-efficiency storage location. High-efficiency storage locations are defined based on shorter average distances to the sorting area. A relocation instruction and the addresses of the old and new storage locations are generated. Optionally, for goods requiring inventory checks, an inventory check instruction is generated and its priority is specified based on their storage duration or special storage requirements, such as storage exceeding 90 days or belonging to a valuable category. The generated inbound, transfer, and inventory check instructions are arranged according to business urgency and execution logic. Business urgency is determined by urgency indicators derived from the order demand spatiotemporal distribution map, and the execution logic follows the dependency of "first transfer to free up storage space, then inbound," forming a storage space allocation instruction sequence.

[0031] In practical implementation, data comparison is reflected in the transformation between the abstract description of the strategy and the specific executable instructions. The "equipment scheduling path instruction set" in the dynamic management strategy set may contain the abstract description of "moving goods from area A to area B", while the "moving equipment control sequence" generated after matching with the map and real-time location contains the precise control command "equipment AGV-001 starts from coordinates (10,20) at 10:00:00, moves along the path point sequence [(12,20),(15,25)], and arrives at coordinates (20,30) at 10:02:30". It can be understood that each instruction in the location allocation instruction sequence is associated with a specific cargo code, source location coordinates, target location coordinates, and expected completion time. The operation node status switching instruction precisely corresponds to the screen prompt or hardware action of the specific workstation. In some embodiments, planning a conflict-free movement path time window requires solving the path intersection and deadlock problem between equipment, which is expressed by the following formula:

[0032] in: Representative equipment The planned entry time for the current path segment Representative equipment The estimated time to complete the previous associated path segment. Representatives and Equipment The set of all other devices that have spatial intersection conflicts in the current path segment. Representative conflict equipment The estimated time to leave the intersection area. This represents a safety time interval reserved to avoid conflicts. Optionally, the preset location selection rule can consider multiple factors when selecting the target location evaluation function, such as the distance from the target location to the frequently used sorting area, the current equipment congestion level in the area, and the frequency of goods storage and retrieval. Finally, the location address with the best comprehensive evaluation is selected and filled into the target address field of the inbound or transfer instruction.

[0033] In one embodiment of the present invention, the deployment of a sensor network and a data acquisition interface is the basis for obtaining feedback after the execution of the control signal queue. The sensor network includes position sensors deployed on the shelf, status sensors installed on the handling equipment, and photoelectric switches at the work nodes. The data acquisition interface is connected to the data output terminals of the warehouse management system and the equipment controller to capture system response data generated during the execution of the control signal queue in real time. The system response data includes the instruction completion status, the actual running trajectory and completion time of the equipment, and the actual occupancy change of the storage location. The instruction completion status is an enumeration value of "success", "failure" or "timeout". The actual running trajectory of the equipment is a coordinate sequence with timestamps. The actual occupancy change of the storage location is an event record of the storage location status changing from "empty" to "occupied" or from "occupied" to "empty". The system response data is compared with the expected execution status in the control signal queue, which includes the expected start point, expected end point, and expected completion time of the instruction. A deviation metric is calculated, reflecting the difference between the preset and actual effects of the strategy. New features reflecting dynamic changes in the warehousing environment and equipment performance drift are extracted from the system response data. These new features may include longer transit times for a certain channel at midday or a decrease in the idle speed of a certain piece of equipment compared to its historical average. This forms a model feedback feature set. The deviation metric and the model feedback feature set are used as training samples to incrementally adjust the parameters of the strategy deduction part in the warehousing and logistics simulation model. These parameters may include equipment speed parameters and task processing time parameters. This reduces the difference between the model deduction results and the actual results. After parameter adjustment, the updated warehousing and logistics simulation model is used in the next round of management strategy generation.

[0034] The expected start and end positions and expected completion time of each instruction are extracted from the control signal queue. The actual start and end positions and actual completion time of each instruction are matched against the system response data using the instruction's unique identifier. The spatial offset between the actual and expected start and end positions of each instruction is calculated, as well as the time deviation between the actual and expected completion times. The spatial offsets and time deviations of all instructions are then aggregated, and a weighted average method is used to calculate the comprehensive deviation metric. It is understood that the deviation metric calculation requires assigning appropriate weights to each type of instruction. For example, for high-precision positioning transfer instructions, the weight of the spatial offset should be set higher; for time-sensitive sorting instructions, the weight of the time deviation should be set higher. An example of calculating the comprehensive deviation metric (D) is as follows:

[0035] in: This represents the overall deviation metric calculated after a control signal queue is executed. This represents the total number of comparable instructions in the control signal queue issued this time. Represents the instruction index. Representing the Spatial offset of an instruction Representing the The time deviation of each instruction. and They are the first The instruction's spatial offset weighting coefficient and time deviation weighting coefficient are preset based on the instruction type and business importance. In some embodiments, the spatial offset... The calculation usually uses the Euclidean distance formula, and the time deviation is... It is the absolute value of the difference between the actual completion time and the expected completion time. See Table 1 for a simplified deviation calculation table: Table 1: Schematic diagram of instruction execution deviation calculation

[0036] Optionally, the process of extracting the model feedback feature set involves feature engineering of the system response data. For example, this includes calculating the average passage speed of equipment in a certain area per unit time, statistically analyzing the probability of failure for specific types of instructions, and identifying sudden patterns of order arrivals. These features, after being vectorized, along with their corresponding deviation metrics, constitute training sample pairs for incremental model learning. In practice, incremental adjustments to the parameters of the strategy deduction part in the warehousing and logistics simulation model can be made using online learning algorithms, such as recursive least squares with a forgetting factor. The deviation metric of the new sample is used as a loss signal to adjust the parameters related to handling speed and task processing delay within the model, making the model's predictions of equipment running time and task consumption time under similar scenarios more closely approximate actual observations in the next deduction. It can be understood that data comparison reflects the difference between the expected state of the control signal queue and the actual state recorded in the system response data; the deviation metric... This difference is quantified, while the model feedback feature set attempts to explain the underlying reasons for the difference.

[0037] See Figure 4 The data presents the dynamic changes of the system's core operational indicators over five consecutive monitoring periods. The warehouse capacity utilization rate, represented by the blue curve, shows a steady upward trend, increasing from 85.0% in the initial period to 90.0% in the fifth period. This reflects the continuous effectiveness of the inventory optimization scheme in dynamically adjusting storage locations and optimizing storage density. The equipment utilization rate, represented by the orange curve, shows slight fluctuations but an overall upward trend, increasing from 77.5% to 83.0%, demonstrating the role of the equipment scheduling path instruction set in optimizing equipment load and reducing empty routes. The order fulfillment rate, represented by the green curve, consistently remained at a high level above 92.5%, eventually climbing to 97.5%, verifying the value of the workflow refactoring strategy in improving order response efficiency and ensuring fulfillment stability. The synergistic optimization trend of these three types of indicators confirms that the warehousing and logistics simulation model in this invention, through dynamic management strategy iteration and updates, can continuously improve the overall operational efficiency of the warehousing system.

[0038] In one embodiment of the present invention, the dynamic strategy triggering mechanism based on order demand is a parallel monitoring and response module embedded in the main management process. This module continuously monitors the incoming outbound order flow data, which is pushed in the form of a real-time message stream. For example, it receives each newly generated order record through a message queue and analyzes the characteristic patterns of new orders in real time. These characteristic patterns include the concentration of goods type combinations, whether the total number of goods in the order exceeds a threshold, the regional concentration of delivery addresses, and the urgency level of the order marking. When the goods type, quantity, or urgency level of a new order constitutes a predetermined characteristic pattern (a predefined set of rules, such as "more than 50 orders accumulated in the same receiving area within 10 minutes" or "orders containing specific promotional categories and marked as expedited"), a strategy regeneration process is triggered. In this process, the new order data is integrated into a structured warehousing characteristic dataset. Specifically, this involves updating the spatiotemporal distribution map of order demand and appending the detailed information of the new orders to the relevant goods file records, forming an updated dataset. Based on the updated dataset, the strategy deduction and analysis of the warehousing logistics simulation model is re-executed to generate a set of dynamic management strategies for new order arrival scenarios. The newly generated set of dynamic management strategies is transformed into an executable control signal queue and sent to the warehouse execution control unit for priority execution.

[0039] In some embodiments, real-time analysis of the characteristic patterns of new orders involves windowing the order flow, for example, using a five-minute time window to aggregate and analyze order features within the window. The process of determining whether a predetermined characteristic pattern exists is typically based on a pattern matching algorithm or rule engine. Optionally, a quantitative determination method is achieved by calculating the pattern matching degree, expressed by the following formula:

[0040] in: Represents the set of new orders Compared to the comprehensive matching score of the predetermined feature patterns, Represents the total number of predefined rules. It is a rule index. It is the first The weight coefficient of each rule, This represents the feature vector extracted in real time from the set of new orders. Representing the Definition of a predefined feature pattern rule, function Calculate eigenvectors With rules The overall match score is calculated and output as a value between 0 and 1. When the threshold is exceeded, a predetermined feature pattern is determined, triggering a strategy regeneration process. This regeneration process can be understood as an interruption-response mechanism; it temporarily suspends or adjusts the original regular strategy execution cycle, prioritizing computational resources for handling urgent strategy optimizations triggered by new order patterns. In practice, integrating new order data into the structured warehouse feature dataset not only updates the data points in the spatiotemporal distribution map of order demand but may also cause changes in the predicted access frequency of related goods in the goods storage lifecycle characteristics, potentially affecting the projected load of equipment in the warehouse equipment performance status matrix.

[0041] In some embodiments, data comparison reflects changes in strategy content before and after the triggering mechanism is activated. For example, before triggering, the system may be executing a routine inventory optimization configuration scheme aimed at balancing warehouse capacity. After receiving a surge of orders targeting the same area, the dynamic strategy triggering mechanism is activated. In the newly generated set of dynamic management strategies, the equipment scheduling path instruction set may be adjusted to prioritize processing orders in that area, and the workflow refactoring strategy may be temporarily adjusted to a batch picking process for orders in that area. Optionally, after the newly generated set of dynamic management strategies is converted into an executable control signal queue, it will be marked as high priority when sent to the warehouse execution control unit. The warehouse execution control unit will adjust the order of its internal execution queue according to the instruction priority to ensure that the instructions related to the new strategy are prioritized for scheduling and execution. It can be understood that the entire dynamic strategy triggering mechanism achieves sensitive perception and rapid response to changes in order flow, enabling the management strategy of the warehousing and logistics system to adapt to real-time changes in business needs, thereby maintaining efficient operation even when order patterns change abruptly.

[0042] See Figure 5The graph displays real-time monitoring results of order feature matching within a 5-minute time window. The three curves in the graph correspond to the matching scores of three predefined rules: Rule 1 (regional concentration), Rule 2 (promotional categories), and Rule 3 (urgent orders), with values ​​ranging from 0 to 1. The dashed line represents the trigger threshold (0.7). When the matching score of any rule exceeds this threshold, the system will trigger a strategy regeneration process. Looking at the time series, in the 09:00-09:05 window, the matching score of Rule 3 is close to 0.8, meeting the trigger condition; in the 09:15-09:20 window, the matching score of Rule 3 exceeds the threshold again and reaches its peak; in the 09:20-09:25 window, the matching scores of Rule 1 and Rule 2 reach 0.81 and 0.65 respectively, with Rule 1 meeting the trigger condition. This multi-rule parallel monitoring mechanism can accurately capture sudden characteristic patterns in the order flow, such as the emergence of regionally concentrated orders, promotional category orders, and urgent orders. By combining quantitative matching degree with threshold triggering, the system can achieve sensitive perception and rapid response to sudden changes in order patterns, thereby dynamically adjusting warehousing and logistics management strategies and ensuring that the system can still operate efficiently when order demand fluctuates.

[0043] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A big data-based intelligent management method for warehousing and logistics, characterized in that, The method includes: Collect raw operational data from the entire warehousing and logistics system. This raw operational data includes real-time information on goods entering the warehouse, monitoring data on the storage status within the warehouse, equipment operation logs, and outbound order flow data. The original operational data is subjected to multi-dimensional cross-fusion processing to generate a structured warehousing feature dataset, which includes cargo storage lifecycle characteristics, warehousing equipment performance status matrix, and order demand spatiotemporal distribution map. Based on the pre-built warehousing and logistics simulation model, strategy deduction and analysis are performed on the structured warehousing feature dataset to generate a set of dynamic management strategies for the warehousing and logistics system. The set of dynamic management strategies includes inventory optimization configuration schemes, equipment scheduling path instruction sets, and work process reconstruction strategies. The set of dynamic management strategies is input into the warehouse execution control unit to generate an executable control signal queue, which includes a location allocation instruction sequence, a handling equipment control sequence, and an operation node status switching instruction. The system monitors the operational feedback data of the warehousing and logistics system after executing the control signal queue in real time, and uses the operational feedback data to iteratively update the parameters of the warehousing and logistics simulation model.

2. The intelligent warehousing and logistics management method based on big data according to claim 1, characterized in that, The process of performing multi-dimensional cross-fusion processing on the original operational data to generate a structured warehouse feature dataset includes: Extract cargo attribute tags and cargo entry time stamps from the real-time cargo entry information to construct a cargo file table with the cargo unique code as the primary key; The storage status monitoring data in the warehouse is associated and mapped with the cargo file table to generate a storage history trajectory chain for each cargo, which includes the current location, storage duration, and environmental parameter records. The equipment operation logs are analyzed to extract the location sequence, energy consumption data and fault warning events of the handling equipment, and to generate a performance status matrix of the warehousing equipment that represents the health and load status of the equipment. The order information in the outbound order flow data is aggregated, and cluster analysis is performed according to the predetermined time granularity and spatial region to draw the spatiotemporal distribution map of the order demand, which reflects the distribution of order volume, goods type and target flow direction. The storage history trajectory chain, the performance status matrix of the warehousing equipment, and the spatiotemporal distribution map of order demand are spatiotemporally aligned and associated with features to form the structured warehousing feature dataset.

3. The intelligent warehousing and logistics management method based on big data according to claim 2, characterized in that, The process involves parsing the equipment operation logs, extracting the location sequence, energy consumption data, and fault warning events of the handling equipment, and generating a performance status matrix of the warehousing equipment that characterizes the equipment's health and load status, including: Identify the movement command records and sensor feedback data of each handling device in the equipment operation log, and construct a device position sequence including timestamps, coordinates, and speeds; Extract energy consumption data of the handling equipment under different operating modes from the log power consumption records, and calculate its energy efficiency index per unit of work. The fault diagnosis rule engine is used to scan the device's operation logs and identify events that meet the characteristics of fault precursors as fault warning events. Based on the regularity of the equipment location sequence, the stability of the energy efficiency index, and the frequency of the fault warning events, a comprehensive health score for each handling device is generated using a weighted scoring model. The overall health score and current load information of all handling equipment are summarized to form a multi-row, multi-column performance status matrix of the warehousing equipment.

4. The intelligent warehousing and logistics management method based on big data according to claim 3, characterized in that, The pre-built warehousing and logistics simulation model performs strategy deduction and analysis on the structured warehousing feature dataset to generate a set of dynamic management strategies for the warehousing and logistics system, including: The structured warehousing feature dataset is used as the initial state parameter and loaded into the warehousing and logistics simulation model; In the warehousing and logistics simulation model, multiple simulation scenarios are set up, and each simulation scenario corresponds to a different set of inventory strategies, equipment scheduling rules and work process assumptions. The warehousing and logistics simulation model is driven to run all simulation scenarios in parallel, simulating the evolution of the warehousing and logistics system's operating status under various strategy assumptions over a period of time in the future. During the simulation, key performance indicators (KPIs) for each simulation scenario are calculated and recorded in real time. These KPIs include warehouse capacity utilization, average order fulfillment time, and overall equipment utilization. By comparing and analyzing the key performance indicator data after the completion of all simulated scenarios, the inventory strategy, equipment scheduling rules and work process assumptions corresponding to the optimal values ​​are selected, and the inventory optimization configuration scheme, equipment scheduling path instruction set and work process reconstruction strategy are integrated to generate the aforementioned inventory optimization configuration scheme, equipment scheduling path instruction set and work process reconstruction strategy.

5. The intelligent warehousing and logistics management method based on big data according to claim 4, characterized in that, The step of inputting the dynamic management strategy set into the warehouse execution control unit to generate an executable control signal queue includes: The inventory optimization configuration scheme is analyzed, and the goods storage suggestions are converted into specific inbound, transfer or inventory count instructions. The target storage location for each instruction is determined based on the current storage location occupancy, forming a storage location allocation instruction sequence. The equipment scheduling path instruction set is matched with the warehouse electronic map and the real-time location of the equipment. Execution equipment is assigned to each scheduling path, and a conflict-free movement path time window is planned to generate the control sequence of the handling equipment. The workflow reconstruction strategy is broken down into the state transition conditions and operation steps of each work node. Based on the current system work progress, the corresponding node state change is triggered to form a work node state switching instruction. According to the preset execution priority and timing dependency, the sequence of storage location allocation instructions, the timing sequence of handling equipment control, and the state switching instructions of operation nodes are sorted and encapsulated to generate the executable control signal queue.

6. The intelligent warehousing and logistics management method based on big data according to claim 5, characterized in that, The process involves analyzing the inventory optimization configuration scheme, converting the goods storage suggestions into specific inbound, transfer, or inventory count instructions, and determining the target storage location for each instruction based on the current storage location occupancy, forming a storage location allocation instruction sequence, including: Identify the recommendations regarding cargo storage density and storage location optimization in the aforementioned inventory optimization configuration scheme; For goods recommended for warehousing, the required storage space is calculated based on the goods attribute tags, and candidate storage locations that meet the storage conditions are selected from the set of available storage locations. The target storage location is selected according to the preset storage location selection rules, and an warehousing instruction is generated. For goods that are recommended to be moved, based on the access frequency in their storage history trajectory chain, they are reassigned from their current inefficient storage location to an efficient storage location, and a move instruction and the new and old storage location addresses are generated. For goods that need to be inventoried, an inventory instruction is generated and the priority of the inventory is specified based on their storage time or special storage requirements. The generated inbound instructions, transfer instructions, and inventory instructions are arranged according to the urgency of the business and the order of execution logic to form the storage location allocation instruction sequence.

7. The intelligent warehousing and logistics management method based on big data according to claim 1, characterized in that, The real-time monitoring of the operational feedback data of the warehousing and logistics system after executing the control signal queue, and the use of the operational feedback data to iteratively update the parameters of the warehousing and logistics simulation model, includes: Deploy a sensor network and data acquisition interface to capture system response data generated during the execution of the control signal queue in real time. The system response data includes the instruction completion status, the actual operating trajectory and completion time of the equipment, and the actual occupancy change of the cargo space. The system response data is compared with the expected execution state in the control signal queue to calculate the deviation metric of the strategy execution. The deviation metric reflects the difference between the preset effect and the actual effect of the strategy. Extract new features from the system response data that reflect dynamic changes in the warehousing environment and equipment performance drift to form a model feedback feature set; Using the deviation metric and the model feedback feature set as training samples, the parameters of the strategy deduction part in the warehousing and logistics simulation model are incrementally adjusted to reduce the difference between the model deduction results and the actual results. After the parameters are adjusted, the updated warehousing and logistics simulation model will be used in the next round of management strategy generation.

8. The intelligent warehousing and logistics management method based on big data according to claim 7, characterized in that, The step of comparing the system response data with the expected execution state in the control signal queue and calculating the deviation metric of strategy execution includes: Extract the expected start and end points and expected completion time of each instruction from the control signal queue; Match the actual start and end points of each instruction with the actual completion time from the system response data; Calculate the spatial offset between the actual execution start point and the expected execution start point, and between the actual end point and the expected end point for each instruction; Calculate the time deviation between the actual completion time and the expected completion time for each instruction; The spatial offsets and temporal deviations of all instructions are aggregated, and a weighted average method is used to calculate a comprehensive deviation metric.

9. The intelligent warehousing and logistics management method based on big data according to claim 1, characterized in that, The method also includes a dynamic strategy triggering mechanism based on order demand: Continuously monitor the incoming outbound order flow data and analyze the characteristic patterns of new orders in real time; When the type, quantity, or urgency of goods in a new order constitutes a predetermined characteristic pattern, the strategy regeneration process is triggered. In the strategy regeneration process, the new order data is integrated into the structured warehouse feature dataset to form an updated dataset; Based on the updated dataset, the strategy deduction and analysis of the warehousing and logistics simulation model are re-executed to generate a set of dynamic management strategies for the arrival of new orders. The newly generated set of dynamic management strategies is transformed into an executable control signal queue and sent to the warehouse execution control unit for priority execution.

10. A big data-based intelligent management system for warehousing and logistics, characterized in that, The system includes a data processing server and a warehouse control gateway. The data processing server exchanges data with the warehouse control gateway. The data processing server stores program instructions and model parameters. When the data processing server executes the program instructions, it implements all the steps of the intelligent warehouse logistics management method based on big data as described in any one of claims 1 to 9.

Images