Goods warehouse-in and warehouse-out management method and management system

Abstract

The invention discloses a cargo warehouse-in and warehouse-out management method and management system, and relates to the technical field of information management. Through dynamic storage position optimization, the warehouse-in and warehouse-out frequency, the associated sales relationship and the storage demand characteristics of cargos are combined with a warehouse three-dimensional space model; and generating a cargo storage location allocation scheme by using a multi-objective optimization algorithm. According to the scheme, the picking time of high-frequency warehouse-in and warehouse-out commodities is remarkably shortened, the average time consumption of single-time commodity picking is reduced, meanwhile, the combined picking error rate of associated commodities is reduced, the problem that in the prior art, storage positions are disjointed with warehouse-in and warehouse-out requirements is solved, through predictive process scheduling, cross congestion of operation channels is effectively avoided, and the working efficiency is improved. Therefore, smooth operation of the in-out warehouse process is realized, the occurrence frequency of congestion events of the in-out warehouse channel is reduced, the delay time is shortened, and the problem of congestion of the in-out warehouse process in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of information management technology, specifically to a method and system for managing the entry and exit of goods. Background Technology

[0002] With the rapid development of trade integration and e-commerce, warehousing and logistics management, as a core link in supply chain operations, has a decisive impact on a company's operating costs and market competitiveness in terms of efficiency and accuracy. Among these processes, goods inbound and outbound management is a key aspect of warehousing activities, directly affecting inventory turnover, order fulfillment, and the overall supply chain's responsiveness and customer satisfaction.

[0003] Existing technologies, such as those with announcement numbers CN112734317A and CN112734317B, propose selecting appropriate storage locations based on the packaging information and attributes of goods to be stored, thereby improving the efficiency of goods entering the warehouse. Another example is the warehouse goods management system disclosed in the invention application patent with announcement number CN116409569A, which uses monitoring equipment to perform cyclical photography and fire detection of goods.

[0004] Combining the above solutions, it can be seen that traditional goods inbound and outbound management methods and systems typically rely on preset static rules and manual operation processes. These methods aim to ensure the orderly progress of goods in receiving, storage, picking, packaging, and shipping through standardized work instructions and fixed resource allocation. However, existing technologies have revealed many limitations when dealing with increasingly complex and changing warehousing environments and customer demands. Specifically, on the one hand, existing storage strategies often fail to fully consider the actual frequency and correlation of goods entering and leaving the warehouse, resulting in high-frequency goods being stored deep in the warehouse and related goods being stored in a scattered manner, significantly increasing the working time of pickers and the error rate of joint picking. On the other hand, traditional inbound and outbound processes usually adopt a fixed manual operation mode, lacking dynamic analysis and scheduling of historical inbound and outbound data, which easily leads to cross-congestion of work channels. Furthermore, in the face of emergency scenarios such as sudden urgent orders, existing systems lack flexible response mechanisms. Multiple manual verification steps are too time-consuming and inefficient. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for managing the entry and exit of goods, which solves the problems existing in the background art.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The first aspect of the present invention provides a method and system for managing the entry and exit of goods, including: A1. acquiring multi-source data related to the entry and exit of goods, including historical entry and exit records, real-time inventory status, goods attribute information, warehouse space layout information, personnel and equipment resource status information, order demand information and emergency event information.

[0007] A2. Integrate and preprocess the multi-source goods entry and exit data to generate a standardized dataset in a unified format.

[0008] A3. Based on the standardized dataset, perform dynamic storage location optimization to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation.

[0009] A4. Based on the standardized dataset, perform predictive process scheduling to generate inbound and outbound operation plans that include adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operation paths.

[0010] A5. Based on the standardized dataset, perform intelligent emergency response processing to generate emergency operation instructions that include emergency order priority enhancement, fast picking routes, or emergency replenishment strategies.

[0011] A6. Monitor the execution of goods entry and exit operations, collect operation performance data, and feed the operation performance data back to the data integration and preprocessing steps to iteratively optimize the algorithm model parameters in the dynamic storage location optimization, the predictive process scheduling, and the intelligent emergency response processing.

[0012] A second aspect of the present invention provides a system for performing the goods entry and exit management method of the present invention, comprising: a data acquisition module for multi-source goods entry and exit related data, including historical entry and exit records, real-time inventory status, goods attribute information, warehouse space layout information, personnel and equipment resource status information, order demand information, and emergency event information.

[0013] The data processing module is used to integrate and preprocess the multi-source goods entry and exit data to generate a standardized dataset in a unified format.

[0014] The storage optimization module is used to dynamically optimize storage locations based on the standardized dataset to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation.

[0015] The process scheduling module is used to perform predictive process scheduling based on the standardized dataset to generate inbound and outbound operation plans that include adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operation paths.

[0016] The emergency response module is used to perform intelligent emergency response processing based on the standardized dataset to generate emergency operation instructions that include emergency order priority enhancement, fast picking routes, or emergency replenishment strategies.

[0017] The feedback learning module is used to monitor the execution of goods entry and exit operations, collect operation performance data, and feed the operation performance data back to the data processing module to iteratively optimize the algorithm model parameters in the storage optimization module, the process scheduling module, and the emergency response module.

[0018] The beneficial effects of this invention are as follows: (1) This invention combines the frequency of goods entering and leaving the warehouse, related sales relationships, and storage demand characteristics with the three-dimensional spatial model of the warehouse through dynamic storage location optimization, and uses a multi-objective optimization algorithm to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation. This scheme significantly shortens the picking time of high-frequency goods entering and leaving the warehouse, reduces the average time spent on a single pickup, and reduces the joint picking error rate of related goods, thereby overcoming the problem of the disconnect between storage location and entry and exit demand in the prior art.

[0019] (2) This invention uses predictive process scheduling to analyze historical peak and off-peak periods for inbound and outbound operations, and combines real-time order demand and resource status to predict future workload and congestion points. A dynamic scheduling algorithm is used to generate inbound and outbound operation plans that include adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operational paths. This effectively avoids cross-congestion in operational channels, ensures smooth operation of the inbound and outbound process, reduces the frequency of congestion events in inbound and outbound channels, shortens delay time, and solves the problem of congestion in inbound and outbound processes in existing technologies.

[0020] (3) This invention, through intelligent emergency response processing, can identify urgent orders or sudden replenishment needs in real time, assign them high priority, plan fast picking routes, or locate backup inventory in real time and generate emergency replenishment strategies. This mechanism reduces the average processing time of sudden urgent orders and effectively improves the fulfillment rate of urgent orders. When temporary replenishment is required, it reduces the replenishment response delay rate, improves the response speed and fulfillment capability in emergency scenarios, and overcomes the problem of delayed response in emergency scenarios in the prior art.

[0021] (4) The present invention also performs in-depth integration and preprocessing of multi-source data through a data processing module, providing a high-quality and highly consistent data foundation for subsequent intelligent decision-making. The feedback learning module realizes the continuous adaptive optimization of the system, enabling the management method and system to continuously iterate and improve according to the actual operation, ensuring its long-term efficient and stable operation. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0024] Figure 2 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Reference Figure 1 As shown, the first aspect of the present invention provides a method and system for managing the entry and exit of goods, including: A1. acquiring multi-source data related to the entry and exit of goods, including historical entry and exit records, real-time inventory status, goods attribute information, warehouse space layout information, personnel and equipment resource status information, order demand information, and emergency event information.

[0027] In a specific embodiment of the present invention, the acquisition of multi-source goods entry and exit related data includes: collecting historical entry and exit records, which include the entry time, exit time, quantity, operator, order number, supplier information and customer information for each batch of goods, acquired through a warehouse management system and an enterprise resource planning system, and stored in time sequence; the operation data includes picking error records, normal picking time, shelving time and handling path, acquired through operation terminal scanning records, IoT device monitoring and manual log entry.

[0028] Collect cargo attribute information, namely static and dynamic attribute data of the cargo. The static attribute data includes the cargo's unique identifier, category, size specifications, weight, storage conditions, and shelf life. It is usually obtained from the product data sheet provided by the supplier or the manual entry system at the time of initial warehousing and stored in the cargo master data file. The dynamic attribute data includes the current inventory, quantity in transit, quantity to be picked, and quantity to be put on the shelf. It is obtained through the real-time updated inventory management system, order management system, and transportation management system interface, and data synchronization and verification are required to ensure the timeliness and accuracy of the data. The cargo attribute information is then summarized.

[0029] It should be noted that the dimensions include length, width, and height, as well as storage conditions such as temperature, humidity, and light protection requirements.

[0030] The warehouse space layout information is obtained, namely the physical environment and spatial layout data of the warehouse. The physical environment data includes real-time temperature, humidity, light intensity and air quality of each area, which is collected in real time by environmental sensors deployed throughout the warehouse. The spatial layout data includes the shelving layout, shelving number, shelf height, load-bearing capacity, and the specific geographical coordinates of each functional area, such as receiving area, shipping area, storage area, picking area and packaging area, as well as the path connections between them, which is obtained through the configuration information of the warehouse management system.

[0031] The system acquires personnel and equipment resource status information, including the number of workers, the functional area where each worker is located, their busy status, and hourly wage standard, through the personnel positioning system interface and the scheduling system. The equipment status information includes equipment type such as forklifts and conveyor belts, the quantity of each equipment type, available time, energy consumption unit price, real-time location and trajectory, through the IoT device monitoring and equipment management system interface.

[0032] Obtain order demand information, namely current order and task planning information. The order information includes order number, goods details, quantity, priority and expected delivery time, which is obtained from the order management system. The task planning information includes inbound plan, outbound plan and replenishment plan, which are obtained from the production planning system, sales forecasting system and warehouse supervisor's manual scheduling system.

[0033] Information on emergencies is acquired, including emergency order notifications, reports of cargo damage during transportation, and system fault warnings. This information is received in real time through a dedicated emergency event reporting interface, anomaly monitoring system alarms, and manual data entry mechanisms. For example, emergency orders are identified by a priority field, and transportation damage reports include photos of the damage and quantity information.

[0034] A2. Integrate and preprocess the multi-source goods entry and exit data to generate a standardized dataset in a unified format.

[0035] In a specific embodiment of the present invention, the integration and preprocessing of the multi-source goods entry and exit related data to generate a standardized dataset in a unified format includes: synchronizing timestamps, converting data formats, and validating data types for the multi-source goods entry and exit related data.

[0036] Outlier detection algorithms are used to identify, correct, or remove outliers in the data. Missing value imputation techniques are used to handle missing values ​​in the data.

[0037] It should be noted that the outlier detection algorithms, such as the Z-score algorithm and the IQR quartile method, determine that the data is outlier when it deviates from the mean by 3 times the standard deviation or exceeds the range of [Q1-1.5IQR, Q3+1.5IQR].

[0038] When static attributes are missing, the default value of the supplier data is used. For example, if the shelf life is not entered, the average value of similar goods is used. When dynamic attributes are missing, linear interpolation is used, that is, it is calculated based on adjacent timestamp data.

[0039] Structured data is normalized or standardized, and unstructured data is feature extracted and encoded to construct a multi-dimensional feature vector containing cargo identification, time dimension, spatial dimension, operational dimension, and resource dimension, forming a unified time-series multi-dimensional dataset.

[0040] A3. Based on the standardized dataset, perform dynamic storage location optimization to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation.

[0041] In a specific embodiment of the present invention, the dynamic storage location optimization includes: analyzing the cargo attribute information, the historical inbound and outbound records, and the order demand information to determine the inbound and outbound frequency, related sales relationships, and storage demand characteristics of each type of cargo. The storage demand characteristics include storage location size, shelf load-bearing requirements, and storage environment requirements.

[0042] By combining the warehouse space layout information and the real-time inventory status, a three-dimensional spatial model of the warehouse is constructed, and the available storage location information is updated in real time.

[0043] Based on the inbound and outbound frequency, related sales relationships, and storage demand characteristics of the goods, a multi-objective optimization algorithm is used to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation. The optimization objective function can be expressed as: in, It is a binary variable, representing the goods. Should it be allocated to a storage location? , Indicates the distance from the shipping area to the location. Pick up The estimated time, Goods The frequency of inbound and outbound operations. In position Pick up Historical error rate Goods Importance coefficient, Goods The distance between it and its related products Goods The strength of its association with related products. Indicates related product groups The strength of the association, It is a related product group The dispersion parameters of goods in the warehouse These are weighting coefficients used to balance the priority between optimizing individual products and aggregating product groups as a whole, such as... The increase indicates a greater emphasis on the centralized storage of related product groups.

[0044] It should be noted that the description of the route from the shipping area to the location... Pick up The estimated time is determined based on warehouse space layout information and operational data from historical inbound and outbound records. Specifically, based on the warehouse's three-dimensional spatial model, the distance from the shipping area to the storage location is first determined. The shortest physical path is calculated, for example, by using the A* algorithm to calculate the straight-line distance or actual passageway length, and then combined with the average moving speed of equipment in historical operation data, such as the average forklift speed of 1.5 m / s, to calculate the distance from the shipping area to the location. Pick up The estimated time.

[0045] The location Pick up The historical error rate is specifically determined based on the operational data in the historical inbound and outbound records, and is specifically counted at the warehouse location. The total number of times item i was picked, plus the number of errors that occurred, divided by the total number of errors, gives the position of the item i. Pick up Historical error rate.

[0046] The goods The importance coefficient is determined based on the product attribute information and order demand information. Specifically, a weighted evaluation score method is used, which weights and sums the unit price, priority, out-of-stock impact feature value in the product attribute information and the frequency of occurrence in urgent orders in the order demand information. This sum is the importance coefficient of product i.

[0047] The goods The association strength between related products is specifically calculated using an association rule algorithm: the co-occurrence frequency between products is statistically analyzed, and the higher the co-occurrence frequency, the greater the association strength value.

[0048] The related product group The dispersion parameter of goods in the warehouse is calculated by taking the average distance between all goods storage locations within the group. The greater the distance, the higher the dispersion. For example, if the average distance between the three goods storage locations in the group is 5 meters, take 0.5. If the average distance is 10 meters, take 0.8.

[0049] Based on the aforementioned cargo storage location allocation scheme, a cargo relocation instruction is generated and issued.

[0050] When new goods are received into the warehouse, a goods storage instruction is generated and issued based on their attributes and the current inventory status.

[0051] In a specific embodiment of the present invention, the multi-objective optimization algorithm is based on a genetic algorithm or a particle swarm optimization algorithm.

[0052] It should be noted that both the genetic algorithm and the particle swarm optimization algorithm use the above-mentioned optimization objective function.

[0053] When generating a solution, the multi-objective optimization algorithm restricts the allocation of goods with a weight exceeding a preset weight threshold to the bottom shelf, and prioritizes the allocation of goods with a shelf life less than or equal to a preset shelf life to the storage location near the shipping area, while also meeting the storage requirements of the goods.

[0054] The minimum picking path cost is determined by calculating the total distance from the picking start point to the target storage location and then to the picking end point.

[0055] The maximum space utilization rate is reflected by the occupancy rate of storage spaces and the uniformity of the distribution of vacant storage spaces.

[0056] The optimal correlation aggregation is determined by calculating the storage location distance of related sales goods.

[0057] A4. Based on the standardized dataset, perform predictive process scheduling to generate inbound and outbound operation plans that include adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operation paths.

[0058] In a specific embodiment of the present invention, the predictive process scheduling includes: analyzing the historical inbound and outbound records to identify the historical peak and off-peak periods of inbound and outbound operations.

[0059] By combining the order demand information with the personnel and equipment resource status information, the volume of inbound and outbound operations and potential congestion points in a specific future time period can be predicted.

[0060] Based on the predicted inbound and outbound operation volume and potential congestion points, the status of personnel and equipment resources, and the historical peak and off-peak periods of inbound and outbound operations, a dynamic scheduling algorithm is used to generate an inbound and outbound operation plan that includes adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operation paths. The optimization objective function can be expressed as: in, It is a binary variable representing the task. Is it within the time window? To process, It is a task In the time window The weight reflects its priority. It is a task In the time window The parameter of processing delay risk, It is a task In the time window Parameters that handle potential congestion risks. Indicates time window The overall congestion index. It is to handle tasks In the time window Resource utilization efficiency It is to handle tasks In the time window The cost.

[0061] The tasks include inbound tasks, outbound tasks, picking tasks, etc.

[0062] It should be noted that the dynamic scheduling algorithm is used to output which time window each person should execute in, and based on the time window allocation results, further detailed task plans are generated. 1. Inbound batch adjustment: Based on the time window allocation, inbound tasks will be concentrated in off-peak windows.

[0063] 2. Outbound wave optimization: Outbound tasks within the same time window are merged into waves based on region and priority.

[0064] 3. Avoid conflict in work routes: Plan conflict-free routes based on the distribution of resources, such as personnel and equipment, within the time window.

[0065] It should be noted that the task mentioned In the time window The weights are specifically set based on task priority, with higher priority tasks having greater weights than lower priority tasks. In the time window The specific calculation method for the delay risk parameter is as follows: the expected delivery time of task t minus the end time of window k yields the delay duration. A risk correction factor is set in conjunction with the predicted inbound and outbound operations of window k. The inbound and outbound operations increase exponentially with the risk correction factor. The delay duration is multiplied by the set risk correction factor to obtain the delay risk parameter.

[0066] The task In the time window The specific calculation method for the potential congestion risk parameter is as follows: assign values ​​based on the task type and the congestion sensitivity of the involved area, and sum these values ​​to obtain the congestion risk parameter. For example, tasks involving large goods often require the use of passageways, resulting in a higher congestion risk, while tasks involving small items have a lower congestion risk. Tasks involving potential congestion points generally have a higher congestion risk, while tasks not involving potential congestion points have a lower congestion risk. The specific calculation method for the overall congestion index representing the time window is as follows: The historical average congestion index of window k is statistically analyzed. The average congestion index includes values ​​from 0 to 1. A congestion index correction factor is set based on the predicted inbound and outbound operation volume of window k. The average congestion index is added to the congestion index correction factor to obtain the overall congestion index of the time window. The predicted inbound and outbound operation volume is directly proportional to the congestion index correction factor. For example, if the inbound and outbound operation volume increases by 50% compared to the historical volume, the congestion index correction factor increases by 0.3.

[0067] The processing task In the time window The resource utilization efficiency is specifically calculated as follows: The task... The duration of device usage divided by the time window The available time of the equipment is used to obtain the equipment utilization rate, and it will participate in the task. The number of people divided by the window The number of available personnel is used to obtain the personnel utilization rate. .

[0068] The processing task In the time window The cost includes the sum of labor costs and equipment costs.

[0069] Based on the inbound and outbound operation plan, inbound, picking, packaging and outbound operation instructions are generated and issued. The instructions include specific execution objects, specific task lists, task order, recommended operation paths and estimated completion times.

[0070] In a specific embodiment of the present invention, the dynamic scheduling algorithm is based on a deep reinforcement learning model or a mixed integer linear programming model.

[0071] It should be noted that the deep reinforcement learning model is suitable for handling complex and real-time changing scheduling needs, such as unconventional dynamic scenarios like sudden emergency order insertions or temporary equipment failures. In this embodiment, it is specifically manifested in the management of goods entering and leaving the warehouse. The deep reinforcement learning model abstracts the warehouse operation scenario into an environment and the scheduling decision into the action of the agent. Through a large number of simulation training, the agent learns how to make decisions in a dynamically changing environment in order to minimize delays, congestion, and costs.

[0072] The mixed-integer linear programming model is suitable for handling scheduling needs with clear rules and relatively stable scenarios, such as adjusting inbound batches and dividing outbound waves within a fixed time period. The mixed-integer linear programming model is a quantitative optimization method that solves the optimal solution by establishing mathematical equations, i.e., the objective function mentioned above. It transforms the scheduling objective into an objective function and the resource constraints into constraints, and calculates the optimal task allocation scheme through a professional solver.

[0073] The two methods mentioned above complement each other, and the specific switching can be adaptively adjusted according to the volatility of the usage scenario.

[0074] The inbound and outbound operation volume prediction utilizes a long short-term memory network or a gated recurrent unit network to train on historical time series data.

[0075] It should be noted that the Long Short-Term Memory Network or Gated Recurrent Unit is a deep learning model specifically designed for processing time series data. It can capture long-term dependencies and has higher prediction accuracy than traditional regression models.

[0076] The conflict avoidance of the operation path is achieved by establishing a dynamic obstacle model on the warehouse map and using an improved A* algorithm or a fast random tree algorithm for path planning and real-time obstacle avoidance.

[0077] It should be noted that establishing a dynamic obstacle model specifically involves marking the location and trajectory of mobile devices on a warehouse map in real time, which are considered as dynamic obstacles. The warehouse map is specifically a three-dimensional electronic map. The improved A* algorithm and the fast random tree algorithm are both path planning algorithms, which are relatively mature in the existing technology and will not be elaborated here.

[0078] A5. Based on the standardized dataset, perform intelligent emergency response processing to generate emergency operation instructions that include emergency order priority enhancement, fast picking routes, or emergency replenishment strategies.

[0079] In a specific embodiment of the present invention, the intelligent emergency response processing includes: identifying urgent orders or sudden replenishment needs.

[0080] In response to the priority of the urgent orders, their priority in the work plan is increased accordingly, and a fast picking path is generated using the shortest path algorithm, namely the improved A* algorithm. The improved A* algorithm specifically introduces dynamic obstacle weights and time weights.

[0081] It should be noted that, based on the priority of emergency orders, the priority of emergency orders in the work plan is mapped according to the mapping relationship between the priority of emergency orders stored in the database and the priority in the work plan.

[0082] In response to the sudden replenishment demand, real-time query of end-to-end inventory information is used to locate backup inventory and generate an emergency replenishment strategy.

[0083] Based on the fast picking path, the emergency replenishment strategy, and the increased order priority, an emergency operation instruction is generated and issued.

[0084] A6. Monitor the execution of goods entry and exit operations, collect operation performance data, and feed the operation performance data back to the data integration and preprocessing steps to iteratively optimize the algorithm model parameters in the dynamic storage location optimization, the predictive process scheduling, and the intelligent emergency response processing.

[0085] Reference Figure 2 As shown, a second aspect of the present invention provides a system for performing the goods entry and exit management method of the present invention, comprising: a data acquisition module for multi-source goods entry and exit related data, including historical entry and exit records, real-time inventory status, goods attribute information, warehouse space layout information, personnel and equipment resource status information, order demand information, and emergency event information.

[0086] The data processing module is used to integrate and preprocess the multi-source goods entry and exit data to generate a standardized dataset in a unified format.

[0087] The storage optimization module is used to dynamically optimize storage locations based on the standardized dataset to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation.

[0088] The process scheduling module is used to perform predictive process scheduling based on the standardized dataset to generate inbound and outbound operation plans that include adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operation paths.

[0089] The emergency response module is used to perform intelligent emergency response processing based on the standardized dataset to generate emergency operation instructions that include emergency order priority enhancement, fast picking routes, or emergency replenishment strategies.

[0090] The feedback learning module is used to monitor the execution of goods entry and exit operations, collect operation performance data, and feed the operation performance data back to the data processing module to iteratively optimize the algorithm model parameters in the storage optimization module, the process scheduling module, and the emergency response module.

[0091] It should be noted that the collected job performance data is first fed back to the data processing module and integrated with the original standardized dataset as input for a new round of optimization. Secondly, based on the deviation of each performance indicator in the job performance data from the expected performance indicator, and combined with the mapping relationship table between the deviation of each performance indicator and the algorithm model parameters stored in the database, the algorithm model parameters for different modules are matched and adjusted accordingly. The establishment of the mapping table between the deviation of each performance indicator and the algorithm model parameters is based on scientific principles. First, based on historical performance data of inbound and outbound operations, such as picking path cost, congestion event frequency, and emergency order processing time, statistical analysis is used to extract typical correspondence patterns between the deviation of different performance indicators and the adjustment of algorithm parameters. At the same time, combined with the mathematical principles of each module's algorithm, such as the balancing logic of weight coefficients in multi-objective optimization and the functional relationship between delay risk and workload in dynamic scheduling, the basic mapping structure is derived by domain experts to ensure that the direction of parameter adjustment conforms to the internal logic of the algorithm. Finally, the feedback learning module continuously collects new operation data to verify and dynamically correct the mapping relationship, eliminate invalid correspondences, and supplement new mappings under special scenarios, so that the table can reflect historical patterns and adapt to real-time scenarios, which conforms to the scientific principles of data-driven, logical deduction, and iterative optimization.

[0092] In a specific embodiment of the present invention, the data acquisition module includes: a radio frequency identification reader, a barcode scanner, a computer vision system, and a three-dimensional laser scanning device.

[0093] The radio frequency identification (RFID) reader is used to read the RFID tags on the goods to obtain the goods' attribute information.

[0094] The barcode scanner is used to scan the barcodes of goods or storage locations for data entry. The computer vision system is used to identify goods, monitor storage location status, and track operational processes.

[0095] The three-dimensional laser scanning equipment is used to obtain warehouse space layout information.

[0096] It should be noted that the models involved in this invention, such as genetic algorithms, particle swarm optimization algorithms, and deep reinforcement learning models, except for the core optimization objective function, have some parameter settings and operational details that are common knowledge in the prior art. Those skilled in the art can set them according to the actual scenario. For example, the conventional parameters of genetic algorithms, in addition to the optimization objective function, include population size, crossover probability, mutation probability, and number of iterations, which are all standardized settings for genetic algorithms. The basic parameters of particle swarm optimization algorithms, in addition to the objective function, include the number of particles, inertia weight, and learning factor. The training parameters of deep reinforcement learning models, such as the specific dimensional division of the state space and action space, learning rate, discount factor, and experience replay pool size, are all mature settings in existing reinforcement learning. The solution configuration of mixed integer linear programming models and the detailed parameters of path planning algorithms are also common knowledge in the prior art. Those skilled in the art can flexibly configure them based on conventional experience and specific scenarios without affecting the implementation and reproduction of the technical solution of this invention.

[0097] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for managing the entry and exit of goods, characterized in that, include: A1. Obtain multi-source data related to the entry and exit of goods, including historical entry and exit records, real-time inventory status, goods attribute information, warehouse space layout information, personnel and equipment resource status information, order demand information, and emergency information; A2. Integrate and preprocess the multi-source goods entry and exit data to generate a standardized dataset in a unified format; A3. Based on the standardized dataset, perform dynamic storage location optimization to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation. A4. Based on the standardized dataset, perform predictive process scheduling to generate inbound and outbound operation plans that include inbound batch adjustment, outbound wave optimization, and operation path conflict avoidance. A5. Based on the standardized dataset, perform intelligent emergency response processing to generate emergency operation instructions that include emergency order priority enhancement, fast picking routes, or emergency replenishment strategies. A6. Monitor the execution of goods entry and exit operations, collect operation performance data, and feed the operation performance data back to the data integration and preprocessing steps to iteratively optimize the algorithm model parameters in the dynamic storage location optimization, the predictive process scheduling, and the intelligent emergency response processing.

2. The method for managing the entry and exit of goods according to claim 1, characterized in that, The acquisition of multi-source goods entry and exit data includes: Collect historical inbound and outbound records. These historical inbound and outbound transaction records include the inbound time, outbound time, quantity, operator, order number, supplier information, and customer information for each batch of goods. They also include operational data such as picking error records, normal picking time, shelving time, and handling routes. Collect cargo attribute information, namely static and dynamic attribute data of the cargo. The static attribute data includes the cargo's unique identifier, category, size specifications, weight, storage conditions and shelf life. The dynamic attribute data includes the current inventory, quantity in transit, quantity to be picked and quantity to be put on the shelf. The cargo attribute information is obtained by summarizing the data. Obtain warehouse space layout information, namely the physical environment and spatial layout data of the warehouse. The physical environment data of the warehouse includes the real-time temperature, humidity, light intensity and air quality of each area. The spatial layout data includes the shelving layout, shelving number, shelf height, load-bearing capacity, and the specific geographical coordinates of each functional area, such as receiving area, shipping area, storage area, picking area and packaging area, as well as the path connection relationship between them. Acquire personnel and equipment resource status information, including the number of workers, the functional area where each worker is located, their busy status, and hourly wage standard. The equipment status information includes equipment type such as forklifts and conveyor belts, the quantity of each equipment type, available time, energy consumption unit price, real-time location and trajectory. Obtain order demand information, namely current order and task planning information. The order information includes order number, goods details, quantity, priority and expected delivery time. The task planning information includes inbound plan, outbound plan and replenishment plan. Obtain information on emergencies, including emergency order notifications, reports of cargo damage during transportation, and system failure warnings.

3. The method for managing the entry and exit of goods according to claim 1, characterized in that, The process of integrating and preprocessing the multi-source goods entry and exit data to generate a standardized dataset in a unified format includes: The data related to the entry and exit of goods from multiple sources are synchronized with timestamps, converted with data formats, and validated with data types. Outlier detection algorithms are used to identify, correct, or remove outliers in the data; missing value imputation techniques are used to handle missing values ​​in the data. Structured data is normalized or standardized, and unstructured data is feature extracted and encoded to construct a multi-dimensional feature vector containing cargo identification, time dimension, spatial dimension, operational dimension, and resource dimension, forming a unified time-series multi-dimensional dataset.

4. The method for managing the entry and exit of goods according to claim 1, characterized in that, The dynamic storage location optimization includes: The product attribute information, the historical inbound and outbound records, and the order demand information are analyzed to determine the inbound and outbound frequency, related sales relationships, and storage demand characteristics of each product. The storage demand characteristics include storage location size, shelf load-bearing requirements, and storage environment requirements. By combining the warehouse space layout information and the real-time inventory status, a three-dimensional space model of the warehouse is constructed, and the available storage location information is updated in real time. Based on the inbound and outbound frequency, related sales relationships, and storage demand characteristics of the goods, a multi-objective optimization algorithm is used to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation. The optimization objective function can be expressed as: in, It is a binary variable, representing the goods. Should it be allocated to a storage location? ; Indicates the distance from the shipping area to the location. Pick up The estimated time; Goods The frequency of inbound and outbound goods; In position Pick up Historical error rate; Goods Importance coefficient; Goods The distance between it and its related products; Goods The strength of its association with related products; Indicates related product groups The strength of the association; It is a related product group The dispersion parameters of goods in the warehouse; These are weighting coefficients; Based on the aforementioned cargo storage location allocation scheme, generate and issue cargo relocation instructions; When new goods are received into the warehouse, a goods storage instruction is generated and issued based on their attributes and the current inventory status.

5. A method for managing the entry and exit of goods according to claim 4, characterized in that, The multi-objective optimization algorithm is based on genetic algorithm or particle swarm optimization algorithm; When generating a solution, the multi-objective optimization algorithm restricts the allocation of goods with a weight exceeding a preset weight threshold to the bottom shelf, and prioritizes the allocation of goods with a shelf life less than or equal to a preset shelf life to the storage location near the shipping area, while also meeting the storage requirements of the goods. The minimum picking path cost is determined by calculating the total distance from the picking start point to the target storage location and then to the picking end point; The maximum space utilization rate is reflected by the storage space occupancy rate and the uniformity of the distribution of vacant storage spaces; The optimal correlation aggregation is determined by calculating the storage location distance of related sales goods.

6. The method for managing the entry and exit of goods according to claim 1, characterized in that, The predictive process scheduling includes: Analyze the historical inbound and outbound records to identify the historical peak and off-peak periods for inbound and outbound operations; By combining the order demand information and the personnel and equipment resource status information, the volume of inbound and outbound operations and potential congestion points in a specific future time period can be predicted. Based on the predicted inbound and outbound operation volume and potential congestion points, the status of personnel and equipment resources, and the historical peak and off-peak periods of inbound and outbound operations, a dynamic scheduling algorithm is used to generate an inbound and outbound operation plan that includes adjustments to inbound batches, optimization of outbound waves, and conflict avoidance of operation paths. The optimization objective function can be expressed as: in, It is a binary variable representing the task. Is it within the time window? Process it; It is a task In the time window The weight reflects its priority; It is a task In the time window Processing delay risk parameters; It is a task In the time window Parameters that handle potential congestion risks; Indicates time window The overall congestion index; It is to handle tasks In the time window Resource utilization efficiency; It is to handle tasks In the time window The cost; Based on the inbound and outbound operation plan, inbound, picking, packaging and outbound operation instructions are generated and issued. The instructions include specific execution objects, specific task lists, task order, recommended operation paths and estimated completion times.

7. A method for managing the entry and exit of goods according to claim 6, characterized in that, The dynamic scheduling algorithm is based on a deep reinforcement learning model or a mixed integer linear programming model; The inbound and outbound operation volume prediction utilizes a long short-term memory network or a gated recurrent unit network to train on historical time series data. The conflict avoidance of the operation path is achieved by establishing a dynamic obstacle model on the warehouse map and using an improved A* algorithm or a fast random tree algorithm for path planning and real-time obstacle avoidance.

8. The method for managing the entry and exit of goods according to claim 1, characterized in that, The intelligent emergency response process includes: Identify urgent orders or sudden replenishment needs; For the priority of the urgent orders, their priority in the work plan is increased accordingly, and the shortest path algorithm, namely the improved A* algorithm, is used to generate a fast picking path; In response to the sudden replenishment demand, real-time query of end-to-end inventory information is used to locate backup inventory and generate an emergency replenishment strategy. Based on the fast picking path, the emergency replenishment strategy, and the increased order priority, an emergency operation instruction is generated and issued.

9. A system for implementing the goods entry and exit management method according to claims 1-8, characterized in that, include: The data acquisition module is used for multi-source goods entry and exit data, including historical entry and exit records, real-time inventory status, goods attribute information, warehouse space layout information, personnel and equipment resource status information, order demand information, and emergency event information; The data processing module is used to integrate and preprocess the multi-source goods entry and exit data to generate a standardized dataset in a unified format. The storage optimization module is used to dynamically optimize storage locations based on the standardized dataset to generate a goods storage location allocation scheme that satisfies the minimum picking path cost, maximum space utilization, and optimal correlation aggregation. The process scheduling module is used to perform predictive process scheduling based on the standardized dataset to generate inbound and outbound operation plans that include inbound batch adjustment, outbound wave optimization, and operation path conflict avoidance. The emergency response module is used to perform intelligent emergency response processing based on the standardized dataset to generate emergency operation instructions that include emergency order priority enhancement, fast picking routes, or emergency replenishment strategies. The feedback learning module is used to monitor the execution of goods entry and exit operations, collect operation performance data, and feed the operation performance data back to the data processing module to iteratively optimize the algorithm model parameters in the storage optimization module, the process scheduling module, and the emergency response module.

10. The goods inbound and outbound management system according to claim 9, characterized in that, The data acquisition module includes: Radio frequency identification (RFID) readers, barcode scanners, computer vision systems, and 3D laser scanning equipment; The radio frequency identification (RFID) reader is used to read RFID tags on goods to obtain goods attribute information; The barcode scanner is used to scan the barcodes of goods or storage locations for data entry; the computer vision system is used to identify goods, monitor storage location status, and the operation process. The three-dimensional laser scanning equipment is used to acquire warehouse space layout information.

Images