Artificial intelligence-based supply chain warehouse resource dynamic allocation system and method

Abstract

The application discloses a supply chain warehouse resource dynamic allocation system and method based on artificial intelligence, relates to the technical field of warehouse management, and comprises a management center, wherein the management center is connected with a warehouse collection module, a resource processing module, a dynamic analysis module and an intelligent deployment module; a virtual visual space is constructed for the supply chain warehouse, warehouse resource data and order task data are collected; warehouse resource data is classified and constructed according to order task data, and a resource category storage library is obtained; the available warehouse of a target user is subjected to capacity visualization according to the resource category storage library, and an alternative resource category capacity graph is obtained; the alternative resource category capacity graph is subjected to demand constraint and sorting through the virtual visual space, a warehouse resource capacity sequence is obtained, the target user is subjected to scheme simulation optimization in the virtual visual space according to the warehouse resource capacity sequence, and an optimal dynamic allocation scheme is obtained; and the warehouse operation efficiency and the space utilization rate are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of warehouse management technology, specifically to a dynamic allocation system and method for supply chain warehouse resources based on artificial intelligence. Background Technology

[0002] The booming development of the e-commerce industry presents unprecedented challenges and opportunities for supply chain management. As a core node connecting production and consumption, the efficiency of warehousing resource allocation directly determines the response speed, service quality, and operating costs of the entire supply chain. Traditional supply chain warehousing resource allocation models mainly rely on static planning or scheduling based on human experience, making it difficult to adapt to the highly dynamic and uncertain market environment.

[0003] Existing supply chain warehousing resource allocation systems largely rely on preset static rules. When faced with unforeseen circumstances, these rules cannot be adjusted in real time, leading to unoptimized locations for popular goods and increased picking distances. Traditional algorithms often optimize single stages, lacking a holistic consideration of the overall warehousing workflow. This may result in efficiency improvements in one stage but cause congestion in downstream stages, reducing overall throughput. Therefore, there is an urgent need to introduce more advanced data processing and intelligent analysis technologies. Breakthroughs in artificial intelligence offer a new technological path to address these pain points. To this end, we present an AI-based dynamic allocation system and method for supply chain warehousing resources. By learning from historical data and sensing real-time status, it predicts order flow and resource demand, dynamically generating optimal resource allocation schemes to achieve real-time, intelligent scheduling of resources such as warehousing space, equipment, and labor. Summary of the Invention

[0004] The objective of this invention can be achieved through the following technical solutions: An AI-based dynamic allocation system for supply chain warehousing resources includes a management center connected to a warehousing data acquisition module, a resource processing module, a dynamic analysis module, and an intelligent deployment module. The warehouse data acquisition module is used to build a virtual visual space for supply chain warehousing and to set up an information acquisition terminal to collect warehouse resource data and target user order task data. The resource processing module is used to extract and classify warehouse resource data based on order task data, obtain resource categories to be allocated, and build a resource category repository. The dynamic analysis module is used to filter the storage availability of target users based on the resource category repository, obtain the available storage points for orders, visualize the resource capacity of the available storage points for orders, and obtain a capacity map of alternative resource categories. The intelligent deployment module is used to constrain the capacity map of candidate resource categories through virtual visual space to obtain a demand resource bounding box, perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints, perform capacity statistics and sorting of available storage points for orders based on the candidate demand endpoints to obtain a storage resource capacity sequence, perform initial task allocation based on virtual visual space to obtain an initial pre-allocation scheme, and optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme.

[0005] Preferably, the process of constructing a virtual visual space includes: Location data is collected from supply chain warehouses to determine storage locations; Construct a virtual visual space, upload the obtained warehouse storage points to the virtual visual space, and perform twin mapping of the warehouse storage points based on the obtained virtual visual space to obtain warehouse twin storage points; The obtained virtual visual space is collected and arranged to obtain an information collection terminal.

[0006] Preferably, the process of collecting warehouse resource data and order task data includes: By collecting resources from the warehouse twin storage points through the information collection terminal, warehouse resource data can be obtained. Time matching is performed on the obtained warehouse resource data to obtain real-time resource collection points; The system collects task data from target users through an information collection terminal to obtain order task data.

[0007] Preferably, the process of obtaining the resource category to be allocated includes: Set the required categories for the activity based on the order task data; Obtain warehouse resource data, classify the warehouse resource data according to the categories required for the activity, and obtain the resource categories to be allocated; Build an initial resource category library, and upload the obtained resource categories to be assigned to the initial resource category library based on the categories required by the activity, thus obtaining the resource category repository.

[0008] Preferably, the process of obtaining a capacity map of alternative resource categories includes: Based on the resource category repository, extract the requirements from the order task data to obtain the order requirement categories; Based on the virtual visual space, the target users are filtered for storage locations according to the obtained order demand categories to obtain available storage locations for orders; Based on the resource category repository, resource coordination is performed on available warehouse points for orders to obtain warehouse storage options; The obtained warehouse reserve resources are converted into capacity to obtain a capacity map of the reserve resource categories.

[0009] Preferably, the process of obtaining the required resource constraint frame includes: Based on the capacity map of alternative resource categories, the order demand categories are visually transformed to obtain a capacity map of demand resource categories. Upload the capacity map of the alternative resource categories to the corresponding order availability storage point in the virtual visual space; Constraints on the capacity map of candidate resource categories are processed using virtual visual space to obtain the constraint boxes of candidate resources; The obtained demand resource category capacity map is uploaded to the virtual visual space, and the demand resource category capacity map is constrained and assimilated by alternative resource constraint boxes to obtain the demand resource bounding boxes.

[0010] Preferably, the process of obtaining the storage resource capacity sequence includes: The original capacity diagram is constructed based on the demand resource constraint box. The candidate resource constraint boxes are concentrically transformed according to the obtained original capacity diagram to obtain the docking capacity diagram. The obtained docking capacity block diagram is marked with poles to obtain alternative demand endpoints; The capacity of the docking capacity diagram is statistically analyzed based on the obtained alternative demand endpoints in the virtual visual space to obtain the comprehensive capacity difference. Based on the obtained comprehensive capacity difference, the available storage points for orders are sorted by quantity to obtain a storage resource capacity sequence.

[0011] Preferably, the process of obtaining the optimal dynamic allocation scheme includes: Based on the warehouse resource capacity sequence, tasks are pre-allocated to target users to obtain an initial pre-allocation plan; The obtained initial pre-allocation plan is uploaded to the virtual visual space. The plan is simulated for the warehouse twin release points corresponding to the available warehouse points of the order through the virtual visual space, and the resource data consumed during the plan simulation is collected and recorded as simulated resource consumption. The obtained simulated resource consumption is uploaded to the docking capacity diagram, and the simulated resource consumption is statistically analyzed for the same type to obtain the difference in simulated resource consumption capacity for each category. The obtained category simulated consumption capacity difference and docking capacity block diagram are sorted by remaining capacity to obtain the storage remaining capacity sequence; The initial pre-allocation scheme is dynamically updated based on the obtained storage capacity sequence to obtain the optimal dynamic allocation scheme.

[0012] Based on the aforementioned AI-based dynamic allocation system for supply chain warehousing resources, this invention also provides an AI-based method for dynamic allocation of supply chain warehousing resources, comprising the following steps: Step 1: Build a virtual visual space for the supply chain warehousing and set up an information collection terminal to collect warehousing resource data and target user order task data. Step 2: Extract and categorize warehouse resource data based on order task data to obtain resource categories to be allocated, and build a resource category repository; Step 3: Filter the storage availability of target users based on the resource category storage repository to obtain available storage points for orders. Visualize the resource capacity of available storage points for orders to obtain a capacity map of alternative resource categories. Step 4: Apply demand constraints to the capacity map of candidate resource categories using virtual visual space to obtain the demand resource bounding box. Perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints. Based on the candidate demand endpoints, perform capacity statistics and sorting of available storage points for orders to obtain a storage resource capacity sequence. Perform initial task allocation based on virtual visual space to obtain an initial pre-allocation scheme. Optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme.

[0013] Compared with the prior art, the beneficial effects of the present invention are: 1. A virtual visual space is built for supply chain warehousing, and the location information of each warehouse is marked to obtain warehouse twin placement points. This establishes a precise mapping of physical warehouses in the digital world, enabling managers to intuitively monitor the overall picture of the warehouse, breaking the information blind spots of physical space, and providing a precise spatial coordinate basis for subsequent intelligent scheduling.

[0014] 2. Collect warehouse resource data from warehouse twin storage locations, and simultaneously collect order task data from target users. Extract resource categories from the warehouse resource data and build a resource category repository. This transforms the originally scattered and unstructured warehouse resource data into standardized data assets, enabling the system to quickly identify and call specific types of resources, thus solving the problem of low matching efficiency caused by data clutter.

[0015] 3. By matching resource categories in order task data, usable warehouse twin storage points for order tasks are obtained and recorded as available warehouse points for orders. This greatly reduces the search space of subsequent optimization algorithms and improves the system's response speed and computational efficiency. The category capacity map of resource category storage is visualized at available warehouse points for orders, and a constraint comparison is made with the category capacity map of order tasks to obtain the comprehensive capacity difference. Complex order demands are quantitatively compared with warehouse carrying capacity, accurately identifying resource gaps or redundancies, and avoiding order delays or resource waste caused by insufficient resource estimation.

[0016] 4. The overall capacity difference is sorted to generate an initial pre-allocation plan. Priority analysis is performed on orders that occur at the same time in the warehouse to obtain a warehouse reserve sequence, ensuring that resources are preferentially allocated to the areas with the best capacity matching, the most scarce resources, or the most favorable conditions. The initial pre-allocation plan is dynamically updated and optimized based on the warehouse reserve sequence to obtain the best dynamic allocation plan. This dynamic feedback mechanism can adapt to unexpected situations such as order changes and equipment failures, eliminate local bottlenecks, and maximize the utilization rate of warehouse resources throughout the entire supply chain. Attached Figure Description

[0017] 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.

[0018] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0020] like Figure 1 As shown, the AI-based supply chain warehousing resource dynamic allocation system includes a management center, which is connected to a warehousing acquisition module, a resource processing module, a dynamic analysis module, and an intelligent deployment module. The warehouse data acquisition module is used to build a virtual visual space for supply chain warehousing and to set up an information acquisition terminal to collect warehouse resource data and target user order task data. The resource processing module is used to extract and classify warehouse resource data based on order task data, obtain resource categories to be allocated, and build a resource category repository. The dynamic analysis module is used to filter the storage availability of target users based on the resource category repository, obtain the available storage points for orders, visualize the resource capacity of the available storage points for orders, and obtain a capacity map of alternative resource categories. The intelligent deployment module is used to constrain the capacity map of candidate resource categories through virtual visual space to obtain a demand resource bounding box, perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints, perform capacity statistics and sorting of available storage points for orders based on the candidate demand endpoints to obtain a storage resource capacity sequence, perform initial task allocation based on virtual visual space to obtain an initial pre-allocation scheme, and optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme.

[0021] In practical applications, although modern warehouse management systems have achieved a certain degree of digitalization, there are still some pain points in resource allocation. For example, traditional allocation often relies on long-term prediction models, information transmission is delayed, and it is difficult to make personalized allocations based on the sales of goods, resulting in low operational efficiency. However, using artificial intelligence technology to dynamically allocate warehouse resources in the supply chain can effectively cope with market demand fluctuations, improve resource utilization, and reduce logistics and transportation costs.

[0022] The warehousing data acquisition module is used to construct a virtual visual space for supply chain warehousing and to set up an information acquisition terminal. This terminal collects warehousing resource data and target user order task data. The specific process includes: Location data is collected from supply chain warehouses to determine storage locations; The location acquisition refers to collecting the warehouse location information of all supply chains that require dynamic resource allocation, obtaining the storage location of the warehouses, which facilitates the construction of a complete supply chain location mapping map and makes it easier to select the most suitable warehouses for resource allocation.

[0023] A virtual visual space is constructed, which is a virtual space used to store virtual models of supply chain warehousing. It can intuitively display the location of warehouses and their positional relationship with adjacent or nearby warehouses.

[0024] The obtained warehouse storage locations are uploaded to the virtual visual space. Based on the supply chain warehousing, the warehouse storage locations are mapped to twins according to the obtained virtual visual space to obtain warehouse twin storage locations. The twin mapping refers to the transformation of the warehouse storage location into a three-dimensional model through virtual visual space, that is, the transformation of the supply chain warehousing into a three-dimensional spatial model of the warehouse storage location, which is denoted as the warehouse twin storage location.

[0025] The obtained virtual visual space is collected and arranged to obtain information collection terminals. The collection arrangement means that according to the location information of the warehouse twin placement points in the virtual visual space, corresponding information collection terminals are set up to collect data information generated by the warehouse twin placement points.

[0026] By collecting resources from the warehouse twin release points through the information collection terminal, warehouse resource data is obtained, and the obtained warehouse resource data is associated with the corresponding warehouse twin release points. It should be further explained that, in the specific implementation process, the warehousing resource data includes physical inventory data, warehousing space resource data, operational resource data, and external environmental data. Among them, physical inventory data represents the information of goods stored in the supply chain warehouse, including but not limited to SKU basic information, dynamic inventory information, and status information. Dynamic inventory data represents the real-time information of each SKU in a specific storage location, such as real-time quantity, batch number, serial number, and production date. Status information represents the current status information of the goods, such as whether the goods are locked, whether they are damaged, or whether they are under quality inspection. Warehouse space resource data includes but is not limited to storage location information, storage location status, and regional topology. Regional topology includes but is not limited to warehouse maps, aisle connectivity, and picking path distance matrices. Operational resource data represents the relevant information of staff and equipment in the warehouse. External environmental data includes but is not limited to warehouse temperature and humidity, local weather conditions, and road congestion.

[0027] The obtained warehousing resource data is time-matched to obtain real-time resource collection points, and the obtained real-time resource collection points are associated with the corresponding warehousing resource data. Furthermore, the time matching means that while collecting warehousing resource data, time information is marked, which is the real-time collection point of the resource, that is, the time point of collection of warehousing resource data, and it is the real-time collection time. In this embodiment, in order to improve the storage accuracy of warehousing goods, it is necessary to improve the time accuracy of time matching, that is, to mark the time to the second, so that the most accurate goods storage status information can be grasped as soon as an order task occurs.

[0028] The system collects task data from target users through an information collection terminal to obtain order task data. The task collection refers to collecting information from target users who need to conduct goods transactions, collecting the target users' order and task data, which is the order task data, including but not limited to order details, work tasks, and time constraints. Among them, order details include SKU + quantity, order type, priority, and customer geographical location; work tasks include tasks to be put on shelves, tasks to be picked, and tasks to be replenished; and time constraints represent the time limit information of the order, such as the order placement time, the promised cut-off time, and the scheduled delivery window.

[0029] The resource processing module is used to extract and classify warehouse resource data based on order task data, obtain resource categories to be allocated, and build a resource category repository. The specific process includes: The activity requires categories based on the obtained order task data. The activity requires categories refer to the types of resources required to complete the order task, which are set according to the order task data. In this embodiment, the activity requires categories are classified according to the following aspects: where the goods are stored, who is responsible for the transportation and storage of the goods, how much inventory is kept for use, and when the activity is carried out.

[0030] Acquire warehousing resource data, classify the acquired warehousing resource data according to the categories required for the activities, obtain the resource categories to be allocated, and associate the obtained resource categories to be allocated with the corresponding activity categories; The resource classification refers to classifying the warehouse resource data of the warehouse twin storage points in the virtual visual space to obtain resource categories that can be allocated, denoted as resource categories to be allocated. In this embodiment, the categories required for the activity are classified according to the following aspects: where the goods are stored, who is responsible for the transportation and storage of the goods, how much inventory is reserved for use, and when the activity will be carried out. The obtained resource categories to be allocated include goods storage resources, work objects, inventory resources, and time resources. Among them, the goods storage resources include storage locations, storage areas, and receiving and shipping platforms; the work objects include human resources and equipment resources; the inventory resources include physical inventory and reserved inventory; and the time resources include work windows and allocation conditions, that is, the time windows in which goods transportation and activities can be arranged.

[0031] An initial resource category library is constructed. Based on the categories required for the activity, the obtained resource categories to be assigned are uploaded to the initial resource category library to obtain the resource category repository. The resource categories to be assigned are uploaded sequentially according to the order of the categories required for the activity, and each resource category is sorted in the repository according to the corresponding real-time resource collection point. That is, in the resource category repository, the resource categories to be assigned are arranged in the order of the categories required for the activity, and each resource category to be assigned can obtain the corresponding category required for the activity.

[0032] The dynamic analysis module is used to filter target users for warehouse storage availability based on resource category repositories, obtain available warehouse storage points for orders, visualize the resource capacity of available warehouse storage points for orders, and obtain a capacity map of candidate resource categories. The specific process includes: Obtain the resource category repository, and extract the requirements from the order task data based on the obtained resource category repository to obtain the order requirement categories; The aforementioned requirement extraction refers to extracting the required categories of activities corresponding to the resource categories to be allocated in the resource category repository from the order task data. The same resource categories are recorded as order requirement categories, indicating which categories of resources the target user's order task requires, such as the name, quantity, pickup time, and delivery location of the required goods. Based on the target user's order requirement categories, the most suitable warehousing resources can be allocated to the target user in the virtual visual space to achieve optimal resource allocation.

[0033] Based on the virtual visual space, the target users are filtered for storage locations according to the obtained order demand categories to obtain available storage locations for orders; It should be further explained that, in the specific implementation process, the storage point screening means that in the virtual visual space, according to the order demand category of each target user, the target user is matched with a warehouse storage point that can carry out the order task delivery, and recorded as an order-available storage point. That is, according to the order quantity and demand time corresponding to the order demand category, the resource category storage point in the warehouse twin storage point is matched with the matching resource category storage point. In other words, the resource category storage point can meet all the resource requirements of the fixed pole demand category. For example, if the order quantity in the resource category storage point is greater than the order quantity of the target user, and the demand time meets the time requirements of the target user, then all resources can meet the order demand of the target user. If any one of them is not met, then the warehouse twin storage point corresponding to the resource category storage point cannot be recorded as an order-available storage point.

[0034] Based on the resource category repository, resource coordination is performed on available warehouse points for orders to obtain warehouse reserve options, and the obtained warehouse reserve options are associated with the corresponding available warehouse points for orders. The resource coordination refers to marking the unallocated resource categories of available warehouse points for orders as warehouse reserve optional resources, which represent the required resources matched when the target user performs an order task. That is, the resources in the available warehouse points corresponding to the order requirement category are marked as warehouse reserve optional resources, which can be used as reserve resources for the target user to complete the order task. Specifically, based on the order demand category, including but not limited to the product name, quantity, pickup time, and delivery location, the corresponding warehouse storage resources, including but not limited to the product name, product inventory quantity, available product quantity, delivery time, and delivery restricted area, vary according to the target user's order demand category and are the available warehouse resources that match the order demand category.

[0035] The obtained warehouse storage options are converted into capacity to obtain a capacity map of the options resource categories. The obtained capacity map of the options resource categories is then associated with the corresponding available warehouse storage points for the orders. It needs further explanation that, in the specific implementation process, the capacity conversion refers to recording the resource categories of the available storage points for each order as a capacity bar, constructing a two-dimensional rectangular coordinate system, and drawing several straight lines outward from the origin of the two-dimensional rectangular coordinate system. The number of straight lines is equal to the number of resource categories of the available storage points. That is, if there are m resource categories of available storage points, then m straight lines are drawn outward from the origin of the two-dimensional rectangular coordinate system. The included angle of each straight line is the same, that is, it evenly divides the 360 ​​degrees of the origin. For example, if there are 6 straight lines, then the included angle between any two adjacent straight lines is 60 degrees. The length of the straight line is determined by the resource quantity of the warehouse's optional resources. That is, how much resource is contained in the warehouse's optional resources is represented by a straight line of corresponding length, which ultimately constitutes a capacity map of alternative resource categories. In the capacity map of alternative resource categories, each type of warehouse optional resources is marked on the corresponding straight line, and the corresponding resource quantity, i.e., the length of the straight line, is displayed. This obtains a display map of the resource quantity of different resource categories in a two-dimensional rectangular coordinate system. It can intuitively obtain a map of the resource capacity changes of available warehouse points for each order, and match warehouses suitable for the target user for resource allocation in the shortest possible time. Under efficient allocation, the order task is completed at the lowest cost.

[0036] The intelligent deployment module is used to constrain the capacity map of candidate resource categories through a virtual visual space to obtain a demand resource bounding box, perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints, sort the available storage points for orders based on the candidate demand endpoints to obtain a storage resource capacity sequence, perform initial task allocation based on the virtual visual space to obtain an initial pre-allocation scheme, and optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme. The specific process includes: Obtain a capacity map of alternative resource categories, and perform visual transformation of order demand categories based on the obtained capacity map of alternative resource categories to obtain a capacity map of demand resource categories; Furthermore, the visual transformation means that based on the obtained candidate resource category capacity map, an initial capacity map of the same form is constructed for the target user's order demand category. The initial capacity map is also represented in a two-dimensional rectangular coordinate system, and the line positions of the resource categories in the constructed initial capacity map are the same as those in the candidate resource category capacity map, except that the line length corresponding to each resource category is not determined. Based on the order demand category, the required resource quantity corresponding to each resource category is marked in the initial capacity map to obtain the demand resource category capacity map.

[0037] Upload the obtained alternative resource category capacity map to the corresponding order availability storage point in the virtual visual space; Constraints on the capacity map of candidate resource categories are processed using virtual visual space to obtain the constraint boxes of candidate resources; It should be further explained that, in the specific implementation process, the constraint processing means that, in the virtual visual space, the endpoints of the straight lines corresponding to the resource capacity of each resource category in the candidate resource category capacity map are marked, and the endpoints of two adjacent straight lines are connected to obtain a closed polygon, denoted as the candidate resource constraint box. Here, the endpoints of the straight lines represent the end endpoints of the lines, that is, the maximum value of the resource capacity. Then, the area of ​​the obtained candidate resource constraint box represents the total resource capacity of the available storage points for the order. Different sides and corners represent the capacity size of that type of resource. By changing the sides and corners of the polygon, the resource capacity change of the corresponding resource category can be obtained, so as to select the most suitable resource for order allocation.

[0038] The obtained demand resource category capacity map is uploaded to the virtual visual space, and the demand resource category capacity map is constrained and assimilated by the alternative resource constraint boxes to obtain the demand resource limit boxes. Furthermore, the constraint assimilation means that, based on the process of obtaining the alternative resource constraint boxes, the demand resource category capacity map is also transformed into a demand resource constraint box of the same form as the alternative resource constraint boxes. The demand resource constraint box is also a closed polygon, and each side and corner represents the resource demand of the corresponding resource category.

[0039] An original capacity diagram is constructed based on the demand resource constraint box. The original capacity diagram is centered on the origin of the demand resource constraint box. The demand resource constraint box is fixed and remains stationary. The other demand resource constraint boxes in the virtual visual space are uploaded to the original capacity diagram with the same origin as the center. The demand resource constraint box and the alternative resource constraint box are marked and displayed in the same diagram to obtain the available storage point for the order that best matches the demand resource constraint box of the target user.

[0040] Based on the obtained original capacity diagram, the candidate resource constraint boxes are concentrically transformed to obtain the docking capacity diagram. The concentric transformation refers to the process of sequentially uploading alternative resource constraint boxes from the origin in the original capacity map to obtain a docking capacity map. During the concentric transformation, the resource categories corresponding to each straight line overlap at the same position, only the line lengths are different, representing different resource capacities and demands. Therefore, in the docking capacity map, the resource difference relationship between the target user's order task's required resources and the available storage resources of the order's warehouse can be directly observed.

[0041] The obtained docking capacity block diagram is marked with poles to obtain alternative demand endpoints, which include alternative extreme points and demand extreme points; It should be further explained that, in the specific implementation process, the extreme point marking means that in the docking capacity block diagram, the end point of each straight line is marked in the demand resource constraint box and the alternative resource constraint box respectively, and the end point is marked as the alternative demand endpoint. Specifically, the end point of each straight line in the demand resource constraint box is marked as the demand extreme point, and the end point of each straight line in the alternative resource constraint box is marked as the alternative extreme point, which is used to intuitively display the maximum capacity of different resource categories.

[0042] The capacity of the docking capacity diagram is statistically analyzed based on the obtained alternative demand endpoints in the virtual visual space to obtain the comprehensive capacity difference, which includes the total demand capacity difference and the category capacity difference. Furthermore, the capacity statistics refer to the calculation of the area difference between each candidate resource constraint box and the original capacity box in the virtual visual space, based on the docking capacity diagram. This difference is recorded as the total demand capacity difference, which is the difference in area between the closed shapes of the candidate resource constraint box and the demand resource limit box. The difference between the candidate extreme point and the demand extreme point of the corresponding resource category line in the candidate resource constraint box and the original capacity box is also calculated and recorded as the category capacity difference. Each resource category corresponding to each line has a corresponding category capacity difference, so each category of resources has a corresponding category capacity difference.

[0043] Based on the obtained comprehensive capacity difference, the available storage points for orders are sorted by content to obtain a storage resource capacity sequence; The content sorting refers to ranking the available storage points for orders in descending order of the comprehensive capacity difference to obtain a storage resource capacity sequence. The ranking is based on the total difference in demand capacity within the comprehensive capacity difference. The higher the ranking of the available storage points for orders, the more resource capacity they have, which can better meet the resource needs of the target user's order tasks. However, a larger resource capacity does not necessarily mean that the target user's order tasks will be completed better. The speed of resource mobilization and allocation within the warehouse also needs to be considered. The ultimate goal is to achieve the most efficient and lowest-cost order task allocation. At this point, it is necessary to consider the degree of demand matching for each type of resource to obtain the most suitable resource allocation scheme.

[0044] Based on the obtained warehouse resource capacity sequence, tasks are pre-allocated to target users to obtain an initial pre-allocation plan; Furthermore, the task pre-allocation refers to generating a preliminary resource allocation scheme for the target user based on the sorting of the warehouse resource capacity sequence, denoted as the initial pre-allocation scheme. For example, according to the sorting of available warehouse points in the warehouse resource capacity sequence, each corresponding warehouse is recorded as an initial pre-allocation scheme. For instance, the first-ranked available warehouse point has an initial pre-allocation scheme. The order is executed according to the warehouse's goods entry and exit rules based on the target user's order task data until the target user's order task is completed. However, in this embodiment, it is only a preliminary fixed allocation scheme for the warehouse. For a warehouse, it may receive order tasks from different target users at the same time. How to arrange the start time of these order tasks is the primary problem to be solved. Therefore, intelligent dynamic allocation is required to meet the order needs of all users. The initial pre-allocation scheme has not obtained the best and most suitable allocation scheme. At this time, it is necessary to simulate the initial pre-allocation scheme in the virtual visual space by combining the category capacity difference of the comprehensive capacity difference. Then, based on the simulation results and the category capacity difference of the comprehensive capacity difference, the initial pre-allocation scheme is optimized and upgraded to obtain the most suitable allocation scheme.

[0045] The obtained initial pre-allocation plan is uploaded to the virtual visual space. The plan is simulated for the warehouse twin release points corresponding to the available warehouse points of the order through the virtual visual space, and the resource data consumed during the plan simulation is collected and recorded as simulated resource consumption. It should be further explained that, in the specific implementation process, the scheme simulation means that in the virtual visual space, the available warehouse points for orders are simulated to execute according to the corresponding initial pre-allocation scheme, that is, the order task is run according to the initial pre-allocation scheme, and the resource data consumed after executing the initial pre-allocation task is counted and recorded as simulated resource consumption, which represents the holding resources of the available warehouse points for orders consumed by the simulated initial pre-allocation scheme. However, in the embodiment, there will not be only one initial pre-allocation scheme at the same time, but multiple initial pre-allocation schemes may appear. Then, the order of the schemes is determined according to the remaining amount of consumed resources, so that the scheme will not have insufficient resources or cannot be completed within the specified time.

[0046] The obtained simulated resource consumption is uploaded to the docking capacity diagram, and the simulated resource consumption is statistically analyzed for the same type to obtain the difference in simulated resource consumption capacity for each category. The aforementioned capacity statistics, as shown in the docking capacity diagram, calculate the category capacity difference, denoted as the category simulated consumption capacity difference. This represents the amount of resources remaining in the warehouse after consuming the resources within the order warehouse in the virtual visual space, following the simulation of the initial pre-allocation scheme.

[0047] The obtained category simulated consumption capacity difference and docking capacity block diagram are sorted by remaining capacity to obtain the storage remaining capacity sequence; The remaining capacity sorting refers to sorting the simulated consumption capacity differences of categories in ascending order to obtain a warehouse remaining capacity sequence. This means that for the same order with available warehouse points, if several target users submit order tasks that need to be completed simultaneously, priority needs to be analyzed to ensure that all order tasks can be completed on time. Based on the warehouse remaining capacity sequence, the order task with the smallest remaining capacity after resource consumption can be executed first. Then, the next resource consumption will be slightly smaller than the previous one, and the replenishment time will be shorter, increasing the speed of order task execution and preventing insufficient inventory.

[0048] The initial pre-allocation plan is dynamically updated based on the obtained storage capacity sequence to obtain the optimal dynamic allocation plan; It should be further explained that, in the specific implementation process, the dynamic update means that, according to the order of the warehouse inventory sequence, the target and user order tasks at the same time are re-allocated. That is, the priority of the scheme is divided according to the order of the warehouse inventory sequence. If there is no time conflict, the order tasks are executed according to the time order. If there is a time conflict, the order is executed according to the order of the warehouse inventory sequence to obtain the final optimal dynamic allocation scheme. That is, the resource allocation within the warehouse is completed. According to the requirements of different order tasks, the final execution order allocation scheme is obtained, which directly limits the flow of resources and realizes intelligent resource mobilization.

[0049] Based on the aforementioned AI-based dynamic allocation system for supply chain warehousing resources, this invention also provides an AI-based method for dynamic allocation of supply chain warehousing resources, comprising the following steps: Step 1: Build a virtual visual space for the supply chain warehousing and set up an information collection terminal to collect warehousing resource data and target user order task data. Step 2: Extract and categorize warehouse resource data based on order task data to obtain resource categories to be allocated, and build a resource category repository; Step 3: Filter the storage availability of target users based on the resource category storage repository to obtain available storage points for orders. Visualize the resource capacity of available storage points for orders to obtain a capacity map of alternative resource categories. Step 4: Apply demand constraints to the capacity map of candidate resource categories using virtual visual space to obtain the demand resource bounding box. Perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints. Based on the candidate demand endpoints, perform capacity statistics and sorting of available storage points for orders to obtain a storage resource capacity sequence. Perform initial task allocation based on virtual visual space to obtain an initial pre-allocation scheme. Optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme.

[0050] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An AI-based dynamic allocation system for supply chain warehousing resources, including a management center, characterized in that: The management center is connected to a warehouse data acquisition module, a resource processing module, a dynamic analysis module, and an intelligent deployment module; The warehouse data acquisition module is used to build a virtual visual space for supply chain warehousing and to set up an information acquisition terminal to collect warehouse resource data and target user order task data. The resource processing module is used to extract and classify warehouse resource data based on order task data, obtain resource categories to be allocated, and build a resource category repository. The dynamic analysis module is used to filter the storage availability of target users based on the resource category repository, obtain the available storage points for orders, visualize the resource capacity of the available storage points for orders, and obtain a capacity map of alternative resource categories. The intelligent deployment module is used to constrain the capacity map of candidate resource categories through virtual visual space to obtain a demand resource bounding box, perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints, perform capacity statistics and sorting of available storage points for orders based on the candidate demand endpoints to obtain a storage resource capacity sequence, perform initial task allocation based on virtual visual space to obtain an initial pre-allocation scheme, and optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme.

2. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 1, characterized in that, The process of constructing a virtual visual space includes: Location data is collected from supply chain warehouses to determine storage locations; Construct a virtual visual space, upload the obtained warehouse storage points to the virtual visual space, and perform twin mapping of the warehouse storage points based on the obtained virtual visual space to obtain warehouse twin storage points; The obtained virtual visual space is collected and arranged to obtain an information collection terminal.

3. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 1, characterized in that, The process of collecting warehouse resource data and order task data includes: By collecting resources from the warehouse twin storage points through the information collection terminal, warehouse resource data can be obtained. Time matching is performed on the obtained warehouse resource data to obtain real-time resource collection points; The system collects task data from target users through an information collection terminal to obtain order task data.

4. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 1, characterized in that, The process of obtaining resource categories to be allocated includes: Set the required categories for the activity based on the order task data; Obtain warehouse resource data, classify the warehouse resource data according to the categories required for the activity, and obtain the resource categories to be allocated; Build an initial resource category library, and upload the obtained resource categories to be assigned to the initial resource category library based on the categories required by the activity, thus obtaining the resource category repository.

5. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 1, characterized in that, The process of obtaining a capacity map of alternative resource categories includes: Based on the resource category repository, extract the requirements from the order task data to obtain the order requirement categories; Based on the virtual visual space, the target users are filtered for storage locations according to the obtained order demand categories to obtain available storage locations for orders; Based on the resource category repository, resource coordination is performed on available warehouse points for orders to obtain warehouse storage options; The obtained warehouse reserve resources are converted into capacity to obtain a capacity map of the reserve resource categories.

6. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 5, characterized in that, The process of obtaining the required resource constraints includes: Based on the capacity map of alternative resource categories, the order demand categories are visually transformed to obtain a capacity map of demand resource categories. Upload the capacity map of the alternative resource categories to the corresponding order availability storage point in the virtual visual space; Constraints on the capacity map of candidate resource categories are processed using virtual visual space to obtain the constraint boxes of candidate resources; The obtained demand resource category capacity map is uploaded to the virtual visual space, and the demand resource category capacity map is constrained and assimilated by alternative resource constraint boxes to obtain the demand resource bounding boxes.

7. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 6, characterized in that, The process of obtaining the storage resource capacity sequence includes: The original capacity diagram is constructed based on the demand resource constraint box. The candidate resource constraint boxes are concentrically transformed according to the obtained original capacity diagram to obtain the docking capacity diagram. The obtained docking capacity block diagram is marked with poles to obtain alternative demand endpoints; The capacity of the docking capacity diagram is statistically analyzed based on the obtained alternative demand endpoints in the virtual visual space to obtain the comprehensive capacity difference. Based on the obtained comprehensive capacity difference, the available storage points for orders are sorted by quantity to obtain a storage resource capacity sequence.

8. The AI-based dynamic allocation system for supply chain warehousing resources according to claim 7, characterized in that, The process of obtaining the optimal dynamic allocation scheme includes: Based on the warehouse resource capacity sequence, tasks are pre-allocated to target users to obtain an initial pre-allocation plan; The obtained initial pre-allocation plan is uploaded to the virtual visual space. The plan is simulated for the warehouse twin release points corresponding to the available warehouse points of the order through the virtual visual space, and the resource data consumed during the plan simulation is collected and recorded as simulated resource consumption. The obtained simulated resource consumption is uploaded to the docking capacity diagram, and the simulated resource consumption is statistically analyzed for the same type to obtain the difference in simulated resource consumption capacity for each category. The obtained category simulated consumption capacity difference and docking capacity block diagram are sorted by remaining capacity to obtain the storage remaining capacity sequence; The initial pre-allocation scheme is dynamically updated based on the obtained storage capacity sequence to obtain the optimal dynamic allocation scheme.

9. The method for dynamic allocation of supply chain warehousing resources in an artificial intelligence-based supply chain warehousing resource dynamic allocation system according to any one of claims 1 to 8, characterized in that, Includes the following steps: Step 1: Build a virtual visual space for the supply chain warehousing and set up an information collection terminal to collect warehousing resource data and target user order task data. Step 2: Extract and categorize warehouse resource data based on order task data to obtain resource categories to be allocated, and build a resource category repository; Step 3: Filter the storage availability of target users based on the resource category storage repository to obtain available storage points for orders. Visualize the resource capacity of available storage points for orders to obtain a capacity map of alternative resource categories. Step 4: Apply demand constraints to the capacity map of candidate resource categories using virtual visual space to obtain the demand resource bounding box. Perform resource category statistics based on the demand resource bounding box to obtain candidate demand endpoints. Based on the candidate demand endpoints, perform capacity statistics and sorting of available storage points for orders to obtain a storage resource capacity sequence. Perform initial task allocation based on virtual visual space to obtain an initial pre-allocation scheme. Optimize the initial pre-allocation scheme based on the storage resource capacity sequence to obtain the optimal dynamic allocation scheme.

Images