A warehouse layout planning method and apparatus

Abstract

The application discloses a warehouse layout planning method and device, and relates to the field of warehouse logistics. A specific embodiment of the method comprises the following steps: obtaining commodity outbound data of a warehouse in a historical production cycle, clustering commodities according to the commodity outbound data to obtain a commodity clustering result; obtaining a plurality of clustering layout schemes according to the commodity clustering result and layout information of a current storage area of the warehouse; determining an optimal clustering layout scheme from the plurality of clustering layout schemes according to model constraint conditions constructed according to a storage area picking scenario; and performing a storage area adjustment operation on commodities in a current production cycle of the warehouse according to the optimal clustering layout scheme. The embodiment can realize intelligent adjustment of commodity storage areas of the warehouse, optimize warehouse space utilization, improve picking efficiency, reduce picking paths and time, and improve the operation efficiency and service level of the entire warehouse management.

Description

Technical Field

[0001] This invention relates to the field of warehousing and logistics, and in particular to a warehouse layout planning method and apparatus. Background Technology

[0002] Currently, warehouse layout planning mainly relies on the experience of warehouse staff; for example, goods arriving at the same time are likely to be placed in similar locations. While this experience-based layout planning method improves operational efficiency to some extent, it also leads to insufficient flexibility and low standardization, hindering the optimization and large-scale development of overall warehouse management. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a warehouse layout planning method and apparatus, which can at least solve the problems of low flexibility and low standardization in existing warehouse layout planning.

[0004] To achieve the above objectives, according to one aspect of the present invention, a warehouse layout planning method is provided, comprising: acquiring goods outbound data of the warehouse during historical production cycles; performing goods clustering based on the goods outbound data to obtain goods clustering results; obtaining multiple clustering layout schemes based on the goods clustering results and the current storage area layout information of the warehouse; constructing model constraints based on the storage area picking scenario; determining the optimal clustering layout scheme from the multiple clustering layout schemes through the model constraints; and performing storage area adjustment operations on goods in the current production cycle of the warehouse according to the optimal clustering layout scheme.

[0005] Optionally, the outbound product data includes order data. The step of clustering products based on the outbound product data to obtain product clustering results includes: determining a set of products in the same order and establishing direct relationships between products in the set; abstracting each product as a city, determining other cities that have direct relationships with each city, and then launching an attack on the other cities; merging cities that failed to be attacked and cities that succeeded to be attacked into one city, and treating the products of the cities that failed to be attacked as appendages of the products of the cities that succeeded to be attacked; generating a city tree structure based on the processed cities; wherein, one city corresponds to one product clustering result.

[0006] Optionally, the process of determining cities that failed to be attacked and cities that succeeded to be attacked includes: identifying a first city and a second city; determining a first military strength value of the first city based on the first commodity outbound data associated with a first commodity in the first city; determining a second military strength value of the second city based on the second commodity outbound data associated with a second commodity in the second city; determining a first advantage value of the first city relative to the second city based on the first commodity outbound data and the second commodity outbound data; determining a first time required for the first city to attack the second city based on the first military strength value and the first advantage value; determining a second time required for the second city to attack the first city based on the second military strength value and the second advantage value; determining the smaller time and the larger time between the first time and the second time, designating the city corresponding to the smaller time as the city that succeeded to be attacked, and designating the city corresponding to the larger time as the city that failed to be attacked.

[0007] Optionally, the method further includes one or more of the following situations: in response to the fact that the failed city has been merged with other successfully attacked cities, the operation of merging the failed city and the successfully attacked city is not performed; in response to the fact that the failed city already has a vassal and the original city corresponding to the vassal is the city of the successfully attacked city, it is determined that there is a cyclic relationship between the failed city, the successfully attacked city and the original city, and the operation of merging the failed city and the successfully attacked city is not performed.

[0008] Optionally, the method further includes: determining the number of cities in the city tree structure and the number of storage areas in the warehouse; adjusting the city tree structure using a corresponding preset city adjustment model based on the relationship between the number of cities and the number of storage areas to obtain a number of cities for the number of storage areas; wherein, the process includes: when the number of cities is greater than the number of storage areas, determining the first target city with the fewest child nodes in the city tree structure, determining other target cities whose lords in the first target city are vassals, merging the first target city and other target cities into one city, repeating the determination of the first target city and the merging operation until a number of cities for the number of storage areas is obtained; when the number of cities is less than the number of storage areas, determining the second target city with the most child nodes in the city tree structure, splitting the vassals in the second target city to obtain multiple new cities, repeating the determination of the second target city and the splitting operation until a number of cities for the number of storage areas is obtained; wherein, the lord of the new city is the lord of the second target city.

[0009] Optionally, the method further includes: in response to the number of first target cities being multiple, determining the power value of the city lord in each first target city based on the quantity of goods in each first target city, and determining the first target city with the smallest power value; in response to the number of second target cities being multiple, determining the power value of the city lord in each second target city based on the quantity of goods in each second target city, and determining the second target city with the largest power value.

[0010] Optionally, obtaining multiple clustering layout schemes based on the product clustering results and the current warehouse storage area layout information includes: using a preset product clustering result scoring model to determine the optimality of each product clustering result; using a preset storage area scoring model to determine the optimality of each storage area; and combining the product clustering results and storage areas in pairs according to the optimality from largest to smallest to obtain multiple clustering layout schemes.

[0011] Optionally, the step of using a preset product clustering result scoring model to determine the optimality of each product clustering result includes: for a single product clustering result, determining the number of products included in the single product clustering result; determining the central product in the single product clustering result, counting the number of orders including the central product in the historical production cycle, and using the counted order number as the popularity of the central product; determining the evaluation popularity of the single product clustering result based on the product outbound data of each product in the single product clustering result; and determining the optimality of the single product clustering result based on one or more of the product quantity, the popularity of the central product, and the evaluation popularity.

[0012] Optionally, determining the evaluation popularity of a single product clustering result based on the outbound data of each product in the single product clustering result includes: determining the average daily inventory and average daily price of each product in the historical production cycle; determining the evaluation popularity of each product based on the average daily inventory and average daily price; and then accumulating these to obtain the total evaluation popularity of all products in the single product clustering result; determining the evaluation popularity of the central product in the single product clustering result; determining the order quantity that includes each product and every other product; using the order quantity as the correlation between each product and every other product; accumulating this to obtain the total correlation between each product and all other products; and then accumulating this to obtain the total correlation between all products in the single product clustering result; determining the total correlation of the central product in the single product clustering result; and determining the evaluation popularity of the single product clustering result based on one or more of the total evaluation popularity of all products, the evaluation popularity of the central product, the total correlation of all products, and the total correlation of the central product.

[0013] Optionally, the step of using a preset storage area scoring model to determine the preference of each storage area includes: for a single storage area, determining the number of vacant storage locations in the single storage area; determining the distance from the single storage area to other storage areas, and then accumulating them to obtain the total distance; determining the distance from the single storage area to the target area; and determining the preference of the single storage area based on one or more of the number of vacant storage locations, the total distance, and the distance from the single storage area to the target area.

[0014] Optionally, determining the optimal clustering layout scheme from the multiple clustering layout schemes based on the model constraints includes: extracting multiple order data from the product outbound data of historical production cycles; for a single clustering layout scheme, determining the storage area where the goods of each order are located, thereby determining the picking distance and picking path of each order, and thus obtaining the total picking path of the multiple order data; determining the target clustering layout scheme where the total picking path is a closed loop, determining the total picking distance of each target clustering layout scheme, and taking the target clustering layout scheme with the smallest total picking distance as the optimal clustering layout scheme.

[0015] To achieve the above objectives, according to another aspect of the present invention, a warehouse layout planning device is provided, comprising: a clustering module, configured to acquire goods outbound data of the warehouse during historical production cycles, and perform goods clustering based on the goods outbound data to obtain goods clustering results; a scheme module, configured to obtain multiple clustering layout schemes based on the goods clustering results and the current storage area layout information of the warehouse; an optimization module, configured to construct model constraints based on the storage area picking scenario, and determine the optimal clustering layout scheme from the multiple clustering layout schemes through the model constraints; and an adjustment module, configured to perform storage area adjustment operations on the goods in the current production cycle of the warehouse according to the optimal clustering layout scheme.

[0016] Optionally, the outbound product data includes order data. The clustering module is used to: determine a set of products in the same order; establish direct relationships between products in the set; abstract each product as a city; determine other cities that have direct relationships with each city; and launch an attack on the other cities; merge cities that fail to be attacked and cities that succeed in being attacked into one city, and treat the products of the cities that fail to be attacked as appendages of the products of the cities that succeed in being attacked; and generate a city tree structure based on the processed cities; wherein, one city corresponds to one product clustering result.

[0017] Optionally, the process of determining cities that failed to be attacked and cities that succeeded to be attacked includes: identifying a first city and a second city; determining a first military strength value of the first city based on the first commodity outbound data associated with a first commodity in the first city; determining a second military strength value of the second city based on the second commodity outbound data associated with a second commodity in the second city; determining a first advantage value of the first city relative to the second city based on the first commodity outbound data and the second commodity outbound data; determining a first time required for the first city to attack the second city based on the first military strength value and the first advantage value; determining a second time required for the second city to attack the first city based on the second military strength value and the second advantage value; determining the smaller time and the larger time between the first time and the second time, designating the city corresponding to the smaller time as the city that succeeded to be attacked, and designating the city corresponding to the larger time as the city that failed to be attacked.

[0018] Optionally, the device further includes one or more of the following situations: in response to the fact that a city that failed to be attacked has been merged with another city that was successfully attacked, the operation of merging the city that failed to be attacked and the city that was successfully attacked is not performed; in response to the fact that a city that failed to be attacked already has a vassal, and the original city corresponding to the vassal is the city that successfully attacked the city, it is determined that there is a cyclic relationship between the city that failed to be attacked, the city that was successfully attacked, and the original city, and the operation of merging the city that failed to be attacked and the city that was successfully attacked is not performed.

[0019] Optionally, the clustering module is further configured to: determine the number of cities in the city tree structure and the number of storage areas in the warehouse; adjust the city tree structure using a corresponding preset city adjustment model according to the relationship between the number of cities and the number of storage areas, to obtain the number of cities in the storage area; wherein, the process includes: when the number of cities is greater than the number of storage areas, determining the first target city with the fewest child nodes in the city tree structure, determining other target cities whose lords are vassals in the first target city, merging the first target city and other target cities into one city, repeating the determination of the first target city and the merging operation until the number of cities in the storage area is obtained; when the number of cities is less than the number of storage areas, determining the second target city with the most child nodes in the city tree structure, splitting the vassals in the second target city to obtain multiple new cities, repeating the determination of the second target city and the splitting operation until the number of cities in the storage area is obtained; wherein, the lord of the new city is the lord of the second target city.

[0020] Optionally, the clustering module is further configured to: in response to the number of first target cities being multiple, determine the power value of the city lord in each first target city based on the quantity of goods in each first target city, and determine the first target city with the smallest power value; in response to the number of second target cities being multiple, determine the power value of the city lord in each second target city based on the quantity of goods in each second target city, and determine the second target city with the largest power value.

[0021] Optionally, the scheme module is used to: use a preset commodity clustering result scoring model to determine the optimization degree of each commodity clustering result; use a preset storage area scoring model to determine the optimization degree of each storage area; and combine the commodity clustering results and storage areas in pairs according to the optimization degree from large to small to obtain multiple clustering layout schemes.

[0022] Optionally, the solution module is configured to: for a single product clustering result, determine the number of products included in the single product clustering result; determine the central product in the single product clustering result, count the number of orders including the central product in the historical production cycle, and use the counted order number as the popularity of the central product; determine the evaluation popularity of the single product clustering result based on the product outbound data of each product in the single product clustering result; and determine the optimization degree of the single product clustering result based on one or more of the product quantity, the popularity of the central product, and the evaluation popularity.

[0023] Optionally, the scheme module is configured to: determine the average daily inventory and average daily price of each product in the historical production cycle; determine the evaluation popularity of each product based on the average daily inventory and average daily price, and then accumulate them to obtain the total evaluation popularity of all products in the single product clustering result; determine the evaluation popularity of the central product in the single product clustering result; determine the order quantity that includes each product and every other product, and use the order quantity as the correlation between each product and every other product, to accumulate the total correlation between each product and all other products, and then accumulate them to obtain the total correlation between all products in the single product clustering result; determine the total correlation of the central product in the single product clustering result; and determine the evaluation popularity of the single product clustering result based on one or more of the total evaluation popularity of all products, the evaluation popularity of the central product, the total correlation of all products, and the total correlation of the central product.

[0024] Optionally, the scheme module is used to: for a single storage area, determine the number of vacant storage locations in the single storage area; determine the distance from the single storage area to other storage areas, and then accumulate them to obtain the total distance; determine the distance from the single storage area to the target area; and determine the preference of the single storage area based on one or more of the number of vacant storage locations, the total distance, and the distance from the single storage area to the target area.

[0025] Optionally, the preferred module is configured to: extract multiple order data from the goods outbound data of historical production cycles; for a single clustering layout scheme, determine the storage area where the goods of each order are located, so as to determine the picking distance and picking path of each order, and thus obtain the total picking path of the multiple order data; determine the target clustering layout scheme with a closed loop for the total picking path, determine the total picking distance of each target clustering layout scheme, and take the target clustering layout scheme with the smallest total picking distance as the optimal clustering layout scheme.

[0026] To achieve the above objectives, according to another aspect of the present invention, an electronic device for warehouse layout planning is provided.

[0027] The electronic device of this invention includes: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement any of the warehouse layout planning methods described above.

[0028] To achieve the above objectives, according to another aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements any of the warehouse layout planning methods described above.

[0029] To achieve the above objectives, according to another aspect of the present invention, a computing program product is provided. One computing program product of the present invention includes a computer program, which, when executed by a processor, implements the warehouse layout planning method provided in the present invention.

[0030] According to the solution provided by the present invention, one embodiment of the above invention has the following advantages or beneficial effects: it can perform cluster analysis on goods based on historical outbound data, generate multiple cluster layout schemes in combination with the current warehouse storage area layout information, and then select the optimal layout scheme according to the model constraints of the storage area picking scenario, thereby realizing intelligent adjustment of the warehouse goods storage area. Its advantages are that it can optimize warehouse space utilization, improve picking efficiency, reduce picking paths and time, and ultimately improve the operational efficiency and service level of the entire warehouse management, solving the problem of low flexibility and standardization caused by the existing reliance on the experience of warehouse personnel.

[0031] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0032] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein:

[0033] Figure 1This is a schematic diagram of the main process of a warehouse layout planning method according to an embodiment of the present invention;

[0034] Figure 2 This is a flowchart illustrating an optional warehouse layout planning method according to an embodiment of the present invention;

[0035] Figure 3(a) is a schematic diagram that abstracts commodities into cities;

[0036] Figure 3(b) is a schematic diagram of determining adjacent cities based on the direct relationship between commodities;

[0037] Figure 3(c) is a schematic diagram of the tree structure of the city;

[0038] Figure 4 This is a flowchart illustrating another optional warehouse layout planning method according to an embodiment of the present invention;

[0039] Figure 5(a) is a schematic diagram of the merged cities;

[0040] Figure 5(b) is a schematic diagram of the divided city;

[0041] Figure 6 This is a flowchart illustrating another optional warehouse layout planning method according to an embodiment of the present invention;

[0042] Figure 7 This is a schematic diagram of the main modules of a warehouse layout planning device according to an embodiment of the present invention;

[0043] Figure 8 This is an exemplary system architecture diagram in which embodiments of the present invention can be applied;

[0044] Figure 9 This is a schematic diagram of the structure of a computer system suitable for implementing the embodiments of the present invention, such as a mobile device or server. Detailed Implementation

[0045] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0046] It should be noted that the collection, gathering, updating, analysis, processing, use, transmission, and storage of user personal information involved in this disclosed technical solution all comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken to prevent unauthorized access to user personal information data and to safeguard user personal information security, network security, and national security.

[0047] This solution applies to both primary and secondary warehouses, but is primarily used for secondary warehouses. Primary warehouses are warehouses or platforms specifically designed for storing and distributing brand-new goods; secondary warehouses, on the other hand, are warehouses or platforms specifically designed for storing and selling secondhand goods, primarily collecting, organizing, and displaying various secondhand items such as furniture, appliances, clothing, and books for consumers to choose from. Therefore, as a branch of reverse business, the secondary business's warehouse operations and characteristics differ significantly from those of typical forward business.

[0048] In terms of business characteristics, the inventory sources for the inbound process include spare parts warehouses, Paipai platform, and return orders from other channels. This means that even for the same SKU (Stock Keeping Unit, the smallest unit used to manage and identify inventory items in an enterprise), the inbound time may differ. Regarding outbound operations, the main business of the second-hand warehouse is currently sales through Paipai, with SKUs primarily focused on cosmetics. A key characteristic is that it generally requires the main product and samples to be bundled together for outbound shipment.

[0049] In terms of business characteristics, the second-hand warehouse introduces the concepts of "second-hand SKU" and "condition". That is, after the same SKU is received into the warehouse, it is graded based on attributes such as its condition, functional integrity, and depreciation rate, generating a corresponding second-hand SKU. All subsequent delisting and circulation operations are based on this second-hand SKU. Therefore, one primary SKU may correspond to multiple second-hand SKUs.

[0050] From an operational perspective, some secondhand warehouses face challenges such as limited space, high staff turnover, and reliance on manual experience for storage area planning. Furthermore, secondhand warehouse operations emphasize daily clearance, meaning orders placed that day must be produced on the same day, placing high demands on timeliness. At the warehouse management system level, the current order processing flow is as follows: after verification and shelving, the system generates picking tasks based on storage location and product information. The allocation logic for picking tasks involves iterating through all available storage locations (including the target SKU) and sequentially allocating them using a greedy strategy. While this allocation method is simple and direct, it can become inefficient when dealing with complex and ever-changing inventory situations.

[0051] Current warehouse layout planning relies heavily on the subjective experience of warehouse staff. This often leads to chaotic shelving and poor warehouse management during handover phases after staff changes or job adjustments. Furthermore, manual layout planning lacks scientific rigor and a macro-level perspective, neglecting long-term planning considerations. Goods arriving at the same time may be placed in adjacent locations, but their categories and departure times are unrelated, making shelving based on arrival time unfounded. This increases the difficulty of picking route planning and wastes manpower and time. When new goods arrive, the existing haphazard planning ultimately perpetuates these errors, further increasing the difficulty of warehouse management and the complexity of warehouse planning.

[0052] Furthermore, existing plans, once finalized, remain unchanged for a long period, lacking adaptability and robustness to changes in the external environment. When external conditions change (such as major sales cycles, temporary promotional activities, or temporary warehouse relocations due to overcrowding in other warehouses), warehouse management often cannot adapt to these adjustments, leading to chaos. In particular, when goods are shipped from secondhand warehouses, the primary method is to bundle main promotional items with the shipment, which is irrelevant to the current layout rules based on the time of receipt. Therefore, a layout planning model suitable for the characteristics of warehouses (such as secondhand warehouses) is needed.

[0053] Furthermore, once a primary SKU enters the secondary warehouse, it is graded based on its condition, generating a new secondary SKU (for example, the same mobile phone, 100% new and 50% new, are two different products in the secondary warehouse). Considering the one-to-many relationship of secondary SKUs, although the inventory in the secondary warehouse is less than that of the primary warehouse, the actual variety of goods managed and circulated in the secondary warehouse far exceeds that of a typical primary warehouse. In addition, with the current integrated primary and secondary warehouse planning, the volume of secondary warehouse business is constantly increasing, necessitating more detailed and scientific layout and management rules.

[0054] See Figure 1 The diagram shows the main flowchart of a warehouse layout planning method provided by an embodiment of the present invention, which includes the following steps:

[0055] S101: Obtain the warehouse's outbound product data during the historical production cycle, perform product clustering based on the outbound product data, and obtain the product clustering results;

[0056] S102: Based on the product clustering results and the current warehouse storage area layout information, multiple clustering layout schemes are obtained;

[0057] S103: Construct model constraints based on the picking scenario in the storage area, and determine the optimal clustering layout scheme from the multiple clustering layout schemes through the model constraints;

[0058] S104: Based on the optimal clustering layout scheme, perform storage area adjustment operations on the goods in the current production cycle of the warehouse.

[0059] In the above implementation, for step S101, the warehouse's outbound product data during historical production cycles is obtained, such as the flow data of ToB (supplier-to-consumer) and ToC (customer-to-consumer) orders in the past six months. This solution performs statistics on a production cycle basis. Assuming the warehouse's daily working hours are 00:00-18:00 and 18:00-24:00, then, based on the production cycle, the daily outbound product data of the previous production cycle, or the daily outbound product data of the previous few production cycles, is obtained to determine the warehouse layout plan for the current or upcoming production cycle. This solution preferably uses outbound product data from recent production cycles, for example, planning the warehouse layout for the second half of the year based on outbound product data from the previous six months.

[0060] It should be noted that the product codes for primary and secondary warehouses differ. For example, the product code for a secondary warehouse consists of the primary SKU code, business type, and grading. The business type indicates the warehousing channel, which can be determined based on the warehousing channel and obtain the code pre-set for the warehousing channel, such as spare parts warehouse (01), Paipai platform (02), and return order (03). The grading is determined based on the condition and appearance of the product, using pre-set grading standards, or a corresponding code can be pre-set for the grade. Taking a mobile phone as an example, suppose there is a secondary mobile phone that, after inspection, has minor scratches on its exterior but is fully functional. According to the warehouse's pre-set grading standards, this mobile phone is rated as 50% new, with a corresponding grading code of 05. Taking a laptop as an example, suppose there is a secondary laptop that is almost brand new, with no scratches on the screen and all functions working normally. According to the warehouse's pre-set grading standards, this laptop is rated as 90% new, with a corresponding grading code of 09. Therefore, assuming a product has the code xxx in the primary warehouse, is stored in the warehouse through the Paipai platform (02) and is in 50% new condition (05), then the product's code in the secondary warehouse is xxx0205. In addition, for secondary warehouses, when generating product codes, it is also necessary to verify whether the product code already exists in the database. For example, if xxx0205 already exists in the database, it does not need to be generated again; otherwise, it needs to be generated a new one.

[0061] After obtaining the warehouse's outbound product data during the historical production cycle, this solution categorizes the goods in the warehouse into multiple product clusters. This process preferably uses the Siege Algorithm, which aims to divide the goods circulating in the warehouse into different "cities" based on the relationships between them. Each city consists of a central product (the lord) and auxiliary products (vassals). It should be noted that the Siege Algorithm is a graph-based classification algorithm. This algorithm can obtain the relationships between different input objects and divide the input objects into multiple clusters.

[0062] In some optimization problems, the process of finding the optimal solution might be vividly described as a "siege," especially when the problem involves breaking through the limitations of existing solutions. However, unlike other existing algorithms, the siege algorithm provided in this solution is characterized by the need to incorporate direct relationships; that is, only connected cities can attack each other. Whether cities are connected depends on whether there is a direct relationship between the goods; that is, only cities built from goods with a direct relationship can be connected.

[0063] For step S102, this scheme establishes a mathematical model for storage area planning based on the siege algorithm. First, based on the product clustering results and the actual layout within the warehouse, multiple clustering layout schemes are obtained.

[0064] First, set the constraints:

[0065] 1) Each storage location holds only one SKU. Although mixed storage is supported, this solution primarily considers the scenario where mixed storage is not allowed. For example, currently, the secondhand warehouse does not support mixed storage, meaning that multiple secondhand SKUs cannot be placed in the same storage location. Therefore, the smallest granularity of inventory management in the secondhand warehouse is the storage location. A single product can be stored in multiple storage locations, and stockouts are not considered.

[0066] 2) Inventory management is systematic, and historical data can be used to guide future placement. Taking second-hand warehouses as an example, because the business of second-hand warehouses is fixed, the types of goods received and outbound orders of each second-hand warehouse within a certain period are systematic, which allows inventory layout to be optimized based on historical data.

[0067] 3) The picking process involves starting from the picking area, proceeding to the shelves to remove the goods from the shelves, and then returning to the picking area to pick, pack, and ship the goods. Furthermore, the picking time for different items is not considered during the picking process, as the actual operation involves directly placing the removed goods into carts or pallet jacks, resulting in minimal time variation. Similarly, the picking action for each shelf takes the same amount of time; therefore, it can be assumed that all picking operations take the same amount of time.

[0068] In one optional implementation, the scheme further includes: determining the target commodity clustering result with the highest optimization degree and determining the target storage area with the highest optimization degree, establishing a correspondence between the target commodity clustering result and the target storage area; and combining the remaining commodity clustering results and remaining storage areas in pairs to obtain multiple clustering layout schemes. The purpose of this scheme is to place the best cities in the optimal storage area, and to place commodities within the same cluster in adjacent storage locations within the storage area. Therefore, the clustering layout scheme = the best cities placed in the optimal storage area + the clustering layout scheme of the remaining cities and remaining storage areas. Therefore, it is necessary to determine the optimization degree of each city (i.e., the commodity clustering result), the optimization degree of each commodity clustering result, and the optimization degree of each storage area. The highest optimization degree is the optimal one; the specific determination results will be discussed later. Figure 6 The description is shown below. Further details will not be provided here. For cities that are not the best and for storage areas that are not optimal, they can be paired up in descending order of preference, or they can be combined arbitrarily. For example, a city with a high preference can be paired with a storage area with a high preference, or vice versa.

[0069] The optimal selection of storage areas is evaluated according to the following criteria: 1) The smaller the average distance from other storage areas, the better; 2) The greater the number of available storage spaces in a storage area, the better, as this allows for the utilization of storage spaces; 3) The smaller the distance from a storage area to the target area (such as the verification packaging area or the picking area), the better. The distance can be calculated using the following methods: the distance between the center points of the areas, the distance between the top left corners of the areas, or other methods, as long as the standard is consistent. This scheme preferentially considers the distance between the center points.

[0070] In addition to the above, this scheme can also consider other methods. For a single storage area, the total distance from that storage area to other storage areas is accumulated, and the storage area's priority is determined by dividing the number of available storage locations by the "total distance multiplied by the distance from the storage area to the target area". For the above three dimensions or two of the three dimensions, weighted summation, priority, scoring, etc., can also be considered to determine the final priority of a single storage area. Taking weighted summation as an example, assuming the weights are: number of available storage locations 0.5, total distance 0.3, distance from the storage area to the target area 0.2, the calculation formula is: Priority = number of available storage locations * 0.5 + total distance from the storage area to other storage areas * 0.3 + distance from the storage area to the target area * 0.2. In this way, the final priority of a single storage area is determined by comprehensively considering various dimensions. However, if only one dimension is considered, the size can be directly compared.

[0071] For steps S103 and S104, model constraints are constructed based on the storage area picking scenario. The merits of each clustering layout scheme are evaluated using these constraints to obtain the optimal clustering layout scheme. Then, based on this scheme, storage area updates are performed on the goods within the warehouse. Inventory management exhibits regularity, and historical data can be used to guide future placement. Taking a used warehouse as an example, because the business of a used warehouse is fixed, the types of goods received and outbound orders within a certain period are regular, allowing inventory layout to be optimized based on historical data. Therefore, the goods within the warehouse also exhibit regularity. Consequently, model constraints can be customized to suit the operational characteristics of used warehouses.

[0072] Model constraints include one or more of the following:

[0073] 1. All goods are placed in storage areas without exceeding the maximum carrying capacity of the storage area. Specifically, the quantity and weight of all goods currently in the warehouse are multiplied together to obtain the weight of each goods, and then summed to obtain the total weight of goods in each city (i.e., the result of goods clustering). Assuming the maximum carrying capacity of the storage area is V1 and the total weight of goods in the city is V2, all storage areas with V2 ≤ V1 are selected.

[0074] 2. Each order's product set is stored in only one storage area. Products shared by multiple orders are stored in multiple storage areas. Taking product clustering results ABC and AEF as an example, this means that A can be stored in two storage areas simultaneously, but ABC needs to be stored in one storage area, and similarly, AEF also needs to be stored in one storage area.

[0075] 3. The picking path is closed-loop and the picking distance is minimized. Multiple order data points are extracted from historical production cycle outbound data. Based on the storage areas where the goods in these orders are located, the picking distance and picking path for each order are determined, thus obtaining the total picking path for completing the multiple orders. A target clustering layout scheme is determined where the total picking path is a closed loop (i.e., starting from the picking area, proceeding to the shelf for de-shelf operation, and then returning to the picking area for picking, packing, and outbound). The total picking distance for each target clustering layout scheme is determined. Based on the total picking distance in ascending order, optional clustering layout schemes are generated, and the optimal clustering layout scheme is determined using the aforementioned two constraints. However, considering only constraint 3, the target clustering layout scheme with the minimum total picking distance can be directly selected as the optimal clustering layout scheme.

[0076] The method for determining the picking distance here needs to be standardized. For example, it can be done according to the following steps:

[0077] 1) Based on the product information in the order, query the storage area location of each product. The storage area location is usually predefined by the warehouse management system, and each storage area has its own coordinates or relative position.

[0078] 2) Based on the storage location of goods in all orders, construct a graph structure where each storage area is a node, and the edges between nodes represent the distance from one storage area to another. If there is a physical connection between storage areas (e.g., a passageway), then establish a weighted edge between the corresponding nodes, with the weight being the actual walking distance between the two storage areas.

[0079] 3) For each order, use a shortest path algorithm (such as Dijkstra's algorithm or A* algorithm) to find the shortest path in the graph structure above, starting from the starting point, passing through the storage areas of all goods in the order, and returning to the starting point. This path is the optimal picking path for the order, and its length is the picking distance.

[0080] Finally, the optimal picking routes for all orders are merged. Taking into account possible duplicate route segments, the entire picking process is optimized to reduce unnecessary walking, thereby obtaining the total picking route for the multiple order data. This process may involve selecting batch picking strategies, such as centralized picking by area or wave picking, to further improve efficiency.

[0081] The method provided in the above embodiments can perform cluster analysis on goods based on historical outbound data, generate multiple cluster layout schemes by combining the current warehouse storage area layout information, and then select the optimal layout scheme according to the model constraints of the storage area picking scenario, thereby realizing intelligent adjustment of the warehouse goods storage area. Its advantages are that it can optimize warehouse space utilization, improve picking efficiency, reduce picking paths and time, and ultimately improve the operational efficiency and service level of the entire warehouse management.

[0082] See Figure 2 The diagram illustrates an optional warehouse layout planning method according to an embodiment of the present invention, comprising the following steps:

[0083] S201: Determine the set of goods in the same order and establish direct associations between the goods in the set;

[0084] S202: Abstract each commodity into a city, identify other cities that have a direct relationship with each city, and then launch an attack on the other cities;

[0085] S203: Merge cities that failed to be attacked and cities that were successfully attacked into one city, and make the goods of the failed city vassals of the goods of the successful city.

[0086] S204: Generate a tree structure of cities based on the processed cities; where one city corresponds to one product clustering result.

[0087] In the above implementation, the siege algorithm is a graph-based classification algorithm. This algorithm can obtain the relationships between different input objects and divide the input objects into multiple clusters. The siege algorithm mainly consists of two steps: attack and adjustment. This scheme mainly describes the attack phase: each object can be abstracted as a city, and all other input objects are considered enemy cities. Each city attacks the other; the victorious side will occupy the defeated side's city, while the defeated side becomes a vassal of the victor, merging with the victor into one city. It should be noted that since the attacks are carried out pairwise sequentially, after each round of attacks, the number of objects in each city will be greater than 1, meaning the total number of objects will increase.

[0088] For steps S201 to S204, in order to avoid the formation of a circular dependency in the division of cities after the attack, the following premises are set: 1) Each city can only attack its adjacent cities; 2) Attacks between cities are carried out in pairs, and all attacks start at the same time; 3) Only adjacent cities can attack each other (which is represented by a line connecting them on the diagram).

[0089] The concept of cities attacking each other does not mean that any two cities can attack each other at will. Rather, it refers to cities with a direct relationship that can attack each other. For example, suppose there are three orders: one order includes goods A, B, and C; another order includes goods B, D, and E; and the third order includes goods C, C, and G. In this case, goods A, B, and C are directly related; goods B, D, and E are directly related; and goods C, C, and G are directly related.

[0090] Taking ABC as an example, A is abstracted as a city, B is abstracted as a city, and C is also abstracted as a city. The same applies to BDE and CFG, ultimately resulting in a total of 7 cities: ABCDEFG, as shown in Figure 3(a). However, only cities B and C have a direct relationship with city A. City B has a direct relationship with cities ACDE, city C has a direct relationship with cities ABFG, city E has a direct relationship with cities BD, city D has a direct relationship with cities BE, city F has a direct relationship with city CG, and city G has a direct relationship with city CF.

[0091] Therefore, when composing the map, cities with direct relationships are placed in adjacent positions. Adjacent positions are not limited to left and right, but also include top and bottom positions, as shown in Figure 3(b). When launching an attack, only other cities with direct relationships are attacked. Taking city A as an example, attacks are launched only against B and C. For B, attacks are launched simultaneously against A, C, D, and E. After these attacks, the above steps result in a tree structure of cities after the attack (a diagram showing the relationship between city lords and vassals), where the number of objects in each city in the tree structure is greater than 1. As shown in Figure 3(c), taking cities AB as an example, if A successfully attacks B, then A and B are merged into one city, and B becomes a vassal of A.

[0092] The process of determining the victor and vanquished is as follows, taking two cities A and B as examples. Let TAB be the time required for A to invade B, VAB be A's advantage over B, and PA be A's military strength. Then, the time A takes to attack B is TAB = 1 / (VAB * PA). Let TBA be the time required for B to invade A, VBA be B's advantage over A, and PB be B's military strength. Then, the time B takes to attack A is TBA = 1 / (VBA * PB). By comparing the values ​​of TAB and TBA, the one with the shorter invasion time is the victor, and the one with the longer invasion time is the vanquished.

[0093] In this regard, taking the force value PA as an example, the calculation method is as follows: determining the number of orders including the commodity A within the historical production cycle, and taking the number of orders as the force value of the A city. Meanwhile, the order quantity is also the heat HA of the commodity A. Taking VAB as an example, the calculation method is: (Heat of Commodity A - Heat of Commodity B) * (Average daily inventory of Commodity A / Average daily inventory of all commodities in the warehouse) * (Average daily quotation of Commodity A / Average daily quotation of Commodity B). For the average inventory of Commodity A, the daily inventory of Commodity A in the historical production cycle is obtained, summed and averaged to obtain the average daily inventory of Commodity A, and the same applies to Commodity B. For the average daily quotation of Commodity A, the daily warehouse quotation of Commodity A in the historical production cycle is obtained, summed and averaged to obtain the average daily quotation of Commodity A. The average daily inventory of all commodities in the warehouse is obtained by summing and averaging the daily inventory of all commodities in the historical production cycle.

[0094] In addition to the comparison of invasion time, when the invasion time is relatively short, there will be a situation where the attack fails. Taking TAB<TBA as an example, two situations will be determined as that A's attack on B fails: 1) City B already has a city owner; 2) The vassal of B can be traced back to A, and the vassals of A and B form a circular dependency. For situation 1), assuming that City C has already occupied City B, then City A cannot occupy B, that is, a city can only be occupied once. For situation 2), assuming that C is a vassal of B, but C can occupy A, then A, B and C form a loop, and it is determined that A's attack on B fails.

[0095] The method provided in the above embodiments abstracts the commodities in the orders as cities and establishes a direct association relationship, simulates the "attack" process to merge the commodities, and finally generates a tree-structured clustering result. The advantages lie in that it can effectively identify and aggregate commodities with close associations, simplify complex order structures, optimize commodity classification and management, and facilitate the efficient execution of subsequent operations such as warehousing, picking and distribution, so as to improve the operational efficiency and service quality of the entire supply chain.

[0096] See Figure 4 , which shows a schematic flowchart of another optional warehouse layout planning method according to an embodiment of the present invention, including the following steps:

[0097] S401: Determine the number of cities in the city tree structure, and determine the number of storage areas in the warehouse;

[0098] S402: According to the magnitude relationship between the number of cities and the number of storage areas, adjust the city tree structure using a corresponding preset city adjustment model, so as to obtain a number of cities equal to the number of storage areas;

[0099] S403: In the case where the number of cities is greater than the number of storage areas, determine a first target city with the smallest number of child nodes in the city tree structure, determine other target cities whose suzerain in the first target city is a vassal, merge the first target city and the other target cities into one city, and repeat the operations of determining the first target city and merging until the number of cities equal to the number of storage areas is obtained;

[0100] S404: In the case where the number of cities is less than the number of storage areas, determine a second target city with the largest number of child nodes in the city tree structure, split the vassals in the second target city to obtain a plurality of new cities, and repeat the operations of determining the second target city and splitting until the number of cities equal to the number of storage areas is obtained; wherein the suzerain of the new cities is the suzerain of the second target city.

[0101] In the above embodiment, the siege algorithm further includes an adjustment step, and this solution mainly describes the adjustment phase: a plurality of cities are generated based on the attack phase, and when an actual problem is processed, the number of cities generated in the attack phase may not match the number of cities actually required. Taking the storage area planning problem as an example, each city at this time corresponds to one storage area, which has actual physical significance. For example, a warehouse with a scale of 700m³ can usually accommodate 25 medium-sized storage areas, so the number of storage areas divided in the attack phase cannot deviate too much from the ideal number of storage areas of the warehouse. Based on this requirement, it is necessary to further adjust the number of cities on the basis of generating a plurality of cities in the attack phase.

[0102] For steps S401 to S404, after all attacks are completed, a suzerain-vassal relationship diagram in a tree structure is obtained, and new cities are generated on this basis, with the purpose of obtaining product groups (each group is a city) equivalent to the required number of storage areas, and each product group has only one central product (suzerain). Assume that the number of storage areas in the warehouse is n, and the number of non-leaf nodes in the city tree structure is m (that is, nodes with vassals). The adjustment steps are performed as follows:

[0103] 1. One node corresponds to one city, and the city tree structure is divided according to each node to obtain m new cities;

[0104] 2. If m=n, the operation is ended; otherwise, a splitting or merging operation is performed, the merging executes step 3, and the splitting executes step 4;

[0105] 3. If m>n, the cities generated in step 1 are merged until m=n and then the operation ends;

[0106] 4. If m<n, the cities generated in step 1 are split until m=n and then the operation ends;

[0107] 5. Through the above operations, obtain a number of cities equal to the number of storage areas. Construct a group of city lords and their associated goods in the cities to obtain the goods clustering results.

[0108] For step 2, find the city with the fewest child nodes from the m cities. Here, a child node is a child node, a child node's child node, a child node's child node's child node, ..., up to a leaf node. A leaf node does not contain any child nodes, meaning it contains 0 child nodes. The number of child nodes in a node's parent-level node is the number of all leaf nodes directly associated with that node. The number of child nodes in a node's parent-level node is the sum of the number of all child nodes directly associated with that node and the number of all leaf nodes directly associated with those child nodes. The goal of this step is to determine the city with the fewest child nodes from all cities.

[0109] Referring to Figure 5(a), assuming that the city CFG contains the fewest child nodes and its lord is C, then identify the other cities associated with C, i.e., the other cities that C is a vassal of, ABC. Merge CFG and ABC to obtain a new city. Then, determine the relationship between (m = m - 1) and n. If they are still not equal, continue the merging operation until m = n.

[0110] For step 3, find the city with the most child nodes from the m cities. Based on the above description, the city with the highest level generally has the most child nodes. Split the vassals in this city to obtain multiple new cities. The number of new cities depends on the number of vassals. Initially, split into two cities, using the principle of equal distribution and rounding down, with vassals randomly assigned to each city. Alternatively, split them one by one, dividing into as many new cities as there are vassals. If m = m + 1 is still not equal to n, continue searching for the city with the most child nodes from the m cities until m = n. Referring to Figure 5(b), if city ABCD has the most child nodes, split the vassal BCD into two parts, B and CD, resulting in two cities. One city includes goods AB, and the other includes A'CD. Here, A' means that the central goods remain unchanged, and goods A can be placed in multiple storage areas.

[0111] Therefore, the principle of splitting and merging in this scheme is: prioritize merging cities with the fewest nodes, and prioritize splitting cities with the most nodes. However, if multiple cities have the same number of nodes, merge the city with the lower lord's power value, and split the city with the higher lord's power value. That is, the power value of each city is determined based on the quantity of its goods; the fewer the goods, the lower the power value, and the more the goods, the higher the power value. Methods for determining power value based on the quantity of goods include, but are not limited to, the following:

[0112] 1. Linear Relationship: Assuming each city has K units of goods, its power value is directly proportional to the number of goods. For example, if a city has 3 units of goods, its power value can be 3; if it has 4 units of goods, its power value can be 4. This method is simple and direct, and the power value increases linearly with the number of goods.

[0113] 2. Exponential Relationship: Another approach is to use an exponential function to determine the power value. For example, assuming the quantity of goods is K, the power value could be 2^K or K^2. This method causes the power value to increase rapidly with the increase in the quantity of goods. For example, when the quantity of goods is 3, the power value is 2^3 = 8; when the quantity of goods is 5, the power value is 2^5 = 32. This non-linear growth method can better reflect the impact of the quantity of goods on the power value.

[0114] The method provided in the above embodiments dynamically adjusts the city-tree structure to match the number of storage areas in the warehouse, ensuring that the product clustering results are adapted to the actual warehouse layout. When the number of cities is greater than or less than the number of storage areas, merging and splitting strategies are adopted respectively to optimize the city structure, ultimately ensuring that each storage area corresponds to a reasonable city. Its advantages lie in its ability to flexibly handle warehouses of different sizes and order structures, ensuring optimal matching between product classification and storage area allocation, thereby improving warehouse space utilization, simplifying management processes, reducing picking paths and time, and enhancing overall warehouse operation efficiency and service levels.

[0115] See Figure 6 The diagram illustrates another optional warehouse layout planning method according to an embodiment of the present invention, including the following steps:

[0116] S601: For a single product clustering result, determine the number of products included in the single product clustering result;

[0117] S602: Determine the central product in the clustering results of a single product, count the number of orders that include the central product during the historical production cycle, and use the counted number of orders as the popularity of the central product;

[0118] S603: Determine the evaluation popularity of a single product clustering result based on the product outbound data of each product in the single product clustering result;

[0119] S604: Determine the optimality of a single product clustering result based on one or more of the following: product quantity, popularity of the central product, and evaluation popularity.

[0120] In the above implementation, storage locations are planned according to the actual situation of the clan and the warehouse. The best cities are placed in the optimal storage area, and goods from the same clan are placed in similar storage locations within the storage area.

[0121] For steps S601 to S604, based on the siege algorithm, multiple clusters centered on core products are obtained. Following a greedy strategy, the best cluster is placed in the best storage area. The best cluster is evaluated based on the following three criteria:

[0122] 1) The cluster contains the most products. This operation can be performed by counting the number of members in each product cluster. Taking the product cluster ABC as an example, the number of members is 3. If we do not consider the center product A, then the number of members is 2. Each product cluster has only one center product, and under the condition of uniform statistical standards, the center product can be ignored.

[0123] 2) Popularity of the central product (i.e., the city lord) within the cluster. Determine the number of orders for the central product within the historical production cycle, and use this number of orders as the popularity of the central product.

[0124] 3) Evaluation popularity of products within a cluster. This can be implemented in several ways. Taking ABC as an example, the correlation between products A, B, and C and other products is determined separately, and then the total correlation of the cluster is accumulated. This total correlation is used as the evaluation popularity. Alternatively, only the correlation of A can be considered, and the correlation of A can be used as the total correlation of the cluster. The correlation is calculated as follows: the correlation between product i and product j is the number of orders containing both product i and product j within the historical production cycle; the total correlation of product i is the sum of the correlations between product i and other products. Here, j is not limited to products in the product clustering results, but corresponds to all products in the historical production cycle.

[0125] In addition to the methods mentioned above, another approach can be considered: multiply the average daily outbound volume of each product over the historical event period by the average daily price, and sum these to obtain the total evaluation popularity of all products in the product clustering result. Alternatively, only consider the average daily outbound volume * average daily price of the central product over the historical event period. Taking samples as an example, they are usually given away as gifts in the primary warehouse, meaning they are not independent products, but they are independent products in the secondary warehouse, so the evaluation popularity of samples also needs to be calculated. A weighted approach can also be considered here. The product clustering result includes multiple products, but the subordinate level is lower than the city lord's level. Therefore, the level of the product clustering result in the city tree structure can be used as the city lord level, and the city lord level minus 1 can be used as the subordinate level, meaning the higher the level, the greater the weight. Using the level as the weight, the evaluation popularity is weighted and summed to obtain the evaluation popularity of the product clustering result.

[0126] One or more of the following can be used as the evaluation popularity of a single product clustering result: the total evaluation popularity of all products, the evaluation popularity of the central product, the total correlation of all products, and the total correlation of the central product. For multiple scenarios, corresponding weights can be pre-set, and a weighted summation method can be used to determine the final evaluation popularity of a single product clustering result.

[0127] The above calculations use three dimensions—product quantity, city owner popularity, and review popularity—to produce three results. Based on one or more of these three results, the optimality of a single product clustering result is determined. The determination methods include, but are not limited to, the following patterns:

[0128] 1. Priority Mode: Set a priority order for each dimension and evaluate them in sequence. For example: if the number of products is the largest, then the product has the highest priority; if the number of products is the same, then compare the popularity of the city lord, the higher the popularity of the city lord, the higher the priority; if the popularity of the city lord is also the same, then compare the popularity of the evaluation, the higher the popularity of the evaluation, the higher the priority.

[0129] 2. Scoring Mode: Scoring criteria are set for each dimension, and the overall score is used to determine the degree of selection. For example, a range of product quantity is set, and the score is determined based on the range. The same applies to the popularity of the city owner and the popularity of the rating. The three are added together to obtain the total score, and the degree of selection is determined based on the total score.

[0130] 3. Weighted Summation Mode: Different weights are assigned to each dimension, and the results are summed to obtain the final optimization score. For example, the weight of product quantity is 0.5, the weight of city lord popularity is 0.3, and the weight of review popularity is 0.2. The calculation formula is: Optimization Score = (Product Quantity * 0.5) + (City Lord Popularity * 0.3) + (Review Popularity * 0.2).

[0131] These models can be flexibly selected or combined according to specific needs to determine the optimality of clustering results for individual products. Three dimensions were considered above; in practice, only one or two dimensions may be considered, and this solution does not restrict the specific consideration method.

[0132] The method provided in the above embodiments quantitatively evaluates the optimality of each product cluster result by comprehensively considering the number of products in a single product cluster, the popularity of the central product, and the overall evaluation popularity. Its advantage lies in its ability to comprehensively reflect the actual needs and importance of product clustering, accurately identify high-value and high-frequency product combinations, thereby providing a scientific basis for warehouse layout optimization, inventory management, and picking strategies, further improving warehouse operational efficiency, reducing operating costs, and enhancing customer service levels.

[0133] This invention introduces an intelligent algorithm for warehouse merchandise storage area planning. This algorithm is integrated with a warehouse management system to improve the scientific nature of storage area planning, thereby shortening the average picking distance and time, increasing picking efficiency, saving labor costs, reducing order fulfillment difficulties, and achieving cost reduction and efficiency improvement. Furthermore, this solution is a closed loop: warehousing-classification-clustering using an algorithm-shelving system-picking-out. Merchandise outbound data is used for clustering using an algorithm to continuously adjust the optimal clustering layout based on current warehouse production conditions and operational characteristics, thus improving the scientific nature of warehouse inventory planning and the robustness of the system.

[0134] See Figure 7 The diagram shows the main modules of a warehouse layout planning device 700 provided in an embodiment of the present invention, including:

[0135] Clustering module 701 is used to obtain the goods outbound data of the warehouse in the historical production cycle, perform goods clustering based on the goods outbound data, and obtain the goods clustering results;

[0136] Scheme module 702 is used to obtain multiple clustering layout schemes based on the product clustering results and the current storage area layout information of the warehouse;

[0137] The optimization module 703 is used to construct model constraints based on the storage area picking scenario, and determine the optimal clustering layout scheme from the multiple clustering layout schemes through the model constraints.

[0138] The adjustment module 704 is used to adjust the storage areas of goods in the current production cycle of the warehouse according to the optimal clustering layout scheme.

[0139] In the implementation device of this invention, the outbound product data includes order data. The clustering module 701 is used for: determining a set of products in the same order; establishing direct relationships between products in the set; abstracting each product as a city; determining other cities that have direct relationships with each city; and launching an attack on the other cities; merging cities that fail to be attacked and cities that succeed in being attacked into one city, and treating the products of the cities that fail to be attacked as appendages of the products of the cities that succeed in being attacked; generating a city tree structure based on the processed cities; wherein, one city corresponds to one product clustering result.

[0140] In the apparatus of this invention, the process of determining cities that failed to be attacked and cities that were successfully attacked includes: determining a first city and a second city; determining a first military strength value of the first city based on the first commodity outbound data associated with a first commodity in the first city; determining a second military strength value of the second city based on the second commodity outbound data associated with a second commodity in the second city; determining a first advantage value of the first city relative to the second city and a second advantage value of the second city relative to the first city based on the first and second commodity outbound data; determining a first time required for the first city to attack the second city based on the first military strength value and the first advantage value, and determining a second time required for the second city to attack the first city based on the second military strength value and the second advantage value; determining the smaller time and the larger time between the first and second times, and designating the city corresponding to the smaller time as a successfully attacked city and the city corresponding to the larger time as a failed attacked city.

[0141] The apparatus of the present invention further includes one or more of the following situations: in response to the fact that a city that failed to be attacked has been merged with other cities that were successfully attacked, the operation of merging the city that failed to be attacked and the city that was successfully attacked is not performed; in response to the fact that a city that failed to be attacked already has a vassal, and the original city corresponding to the vassal is the city of the city that was successfully attacked, it is determined that there is a cyclic relationship between the city that failed to be attacked, the city that was successfully attacked, and the original city, and the operation of merging the city that failed to be attacked and the city that was successfully attacked is not performed.

[0142] In the implementation device of this invention, the clustering module 701 is further used to: determine the number of cities in the city tree structure and the number of storage areas in the warehouse; adjust the city tree structure using a corresponding preset city adjustment model according to the relationship between the number of cities and the number of storage areas to obtain the number of cities in the storage area; wherein, the process includes: when the number of cities is greater than the number of storage areas, determining the first target city with the fewest child nodes in the city tree structure, determining other target cities whose lords are vassals in the first target city, merging the first target city and other target cities into one city, repeating the determination of the first target city and the merging operation until the number of cities in the storage area is obtained; when the number of cities is less than the number of storage areas, determining the second target city with the most child nodes in the city tree structure, splitting the vassals in the second target city to obtain multiple new cities, repeating the determination of the second target city and the splitting operation until the number of cities in the storage area is obtained; wherein, the lord of the new city is the lord of the second target city.

[0143] In the apparatus of this invention, the clustering module 701 is further configured to: in response to the number of first target cities being multiple, determine the power value of the city lord in each first target city based on the quantity of goods in each first target city, and determine the first target city with the smallest power value; in response to the number of second target cities being multiple, determine the power value of the city lord in each second target city based on the quantity of goods in each second target city, and determine the second target city with the largest power value.

[0144] In the implementation device of the present invention, the scheme module 702 is used to: determine the optimization degree of each commodity clustering result using a preset commodity clustering result scoring model; determine the optimization degree of each storage area using a preset storage area scoring model; and combine the commodity clustering results and storage areas in pairs according to the order of optimization degree from large to small to obtain a variety of clustering layout schemes.

[0145] In the implementation device of the present invention, the scheme module 702 is used to: determine the number of products included in a single product clustering result; determine the central product in the single product clustering result, count the number of orders including the central product in the historical production cycle, and use the counted number of orders as the popularity of the central product; determine the evaluation popularity of the single product clustering result based on the product outbound data of each product in the single product clustering result; and determine the optimization degree of the single product clustering result based on one or more of the product quantity, the popularity of the central product, and the evaluation popularity.

[0146] In the implementation device of this invention, the scheme module 702 is used to: determine the average daily inventory and average daily price of each commodity in the historical production cycle; determine the evaluation popularity of each commodity based on the average daily inventory and average daily price, and then accumulate them to obtain the total evaluation popularity of all commodities in a single commodity cluster result; determine the evaluation popularity of the central commodity in the single commodity cluster result; determine the order quantity that includes each commodity and every other commodity, use the order quantity as the correlation between each commodity and every other commodity, and accumulate them to obtain the total correlation between each commodity and all other commodities, and then accumulate them to obtain the total correlation between all commodities in the single commodity cluster result; determine the total correlation of the central commodity in the single commodity cluster result; and determine the evaluation popularity of the single commodity cluster result based on one or more of the total evaluation popularity of all commodities, the evaluation popularity of the central commodity, the total correlation of all commodities, and the total correlation of the central commodity.

[0147] In the implementation device of the present invention, the scheme module 702 is used to: for a single storage area, determine the number of vacant storage spaces in the single storage area; determine the distance from the single storage area to other storage areas, and then accumulate them to obtain the total distance; determine the distance from the single storage area to the target area; and determine the preference of the single storage area based on one or more of the number of vacant storage spaces, the total distance, and the distance from the single storage area to the target area.

[0148] In the apparatus of this invention, the preferred module 703 is used to: extract multiple order data from the commodity outbound data of historical production cycles; for a single clustering layout scheme, determine the storage area where the commodity of each order is located, so as to determine the picking distance and picking path of each order, and then obtain the total picking path of the multiple order data; determine the target clustering layout scheme with a closed loop for the total picking path, determine the total picking distance of each target clustering layout scheme, and take the target clustering layout scheme with the smallest total picking distance as the optimal clustering layout scheme.

[0149] Furthermore, the specific implementation details of the device described in the embodiments of the present invention have been described in detail in the above-described method, so the details will not be repeated here.

[0150] Figure 8 An exemplary system architecture 800 to which embodiments of the present invention can be applied is shown, including terminal devices 801, 802, 803, network 804, and server 805 (only an example).

[0151] Terminal devices 801, 802, and 803 can be various electronic devices with displays and web browsing capabilities, and can be equipped with various communication client applications. Users can use terminal devices 801, 802, and 803 to interact with server 805 via network 804 to receive or send messages, etc.

[0152] Network 804 is a medium used to provide a communication link between terminal devices 801, 802, 803 and server 805. Network 804 can include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0153] Server 805 can be a server providing various services, such as a backend management server supporting shopping websites browsed by users using terminal devices 801, 802, and 803 (this is just an example). The backend management server can analyze and process data such as received product information query requests, and feed back the processing results (such as target push information and product information—this is just an example) to the terminal devices. It should be noted that the method provided in this embodiment of the invention is generally executed by server 805, and correspondingly, the apparatus is generally set in server 805.

[0154] It should be understood that Figure 8The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0155] The following is for reference. Figure 9 It shows a schematic diagram of the structure of a computer system 900 suitable for implementing a terminal device of the present invention. Figure 9 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0156] like Figure 9 As shown, the computer system 900 includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 902 or programs loaded from storage section 908 into random access memory (RAM) 903. The RAM 903 also stores various programs and data required for the operation of the system 900. The CPU 901, ROM 902, and RAM 903 are interconnected via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0157] The following components are connected to I / O interface 905: an input section 906 including a keyboard, mouse, etc.; an output section 907 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN card, modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to I / O interface 905 as needed. A removable medium 911, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 910 as needed so that computer programs read from it can be installed into storage section 908 as needed.

[0158] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 909, and / or installed from removable medium 911. When the computer program is executed by central processing unit (CPU) 901, it performs the functions defined above in the system of this invention.

[0159] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0160] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0161] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be housed in a processor; for example, a processor can be described as including a clustering module, a scheme module, a selection module, and an adjustment module. The names of these modules do not necessarily limit the module itself; for example, the selection module can also be described as a "screening module."

[0162] In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs that, when executed by the device, cause the device to perform any of the warehouse layout planning methods described above.

[0163] The computer program product of the present invention includes a computer program that, when executed by a processor, implements the warehouse layout planning method in the embodiments of the present invention.

[0164] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A warehouse layout planning method, characterized in that, include: Obtain the warehouse's outbound product data during historical production cycles, and perform product clustering based on the outbound product data to obtain product clustering results; Based on the product clustering results and the current warehouse storage area layout information, multiple clustering layout schemes are obtained; Based on the picking scenario in the storage area, model constraints are constructed, and the optimal clustering layout scheme is determined from the multiple clustering layout schemes using the model constraints. Based on the optimal clustering layout scheme, the storage areas of goods within the current production cycle of the warehouse are adjusted.

2. The method according to claim 1, characterized in that, The product outbound data includes order data. The step of clustering products based on the product outbound data to obtain product clustering results includes: Identify the set of products in the same order and establish direct relationships between the products in the set; Each commodity is abstracted into a city, other cities that have a direct relationship with each city are identified, and then attacks are launched against these other cities. Merge cities that failed to be attacked and cities that were successfully attacked into one city, and make the goods of the failed city vassals of the goods of the successful city. A tree structure of cities is generated based on the processed cities; one city corresponds to one product clustering result.

3. The method according to claim 2, characterized in that, The process of identifying cities that failed to be attacked and those that were successfully attacked includes: The first city and the second city are identified. Based on the first commodity outbound data associated with the first commodity in the first city, the first military strength of the first city is determined. Based on the second commodity outbound data associated with the second commodity in the second city, the second military strength of the second city is determined. Based on the first commodity outbound data and the second commodity outbound data, determine the first advantage value of the first city relative to the second city, and determine the second advantage value of the second city relative to the first city. Based on the first strength value and the first advantage value, determine the first time required for the first city to attack the second city, and based on the second strength value and the second advantage value, determine the second time required for the second city to attack the first city. Determine the smaller and larger time periods between the first and second time periods. The city corresponding to the smaller time period is considered a successfully attacked city, and the city corresponding to the larger time period is considered a failed attacked city.

4. The method according to claim 2, characterized in that, The method also includes one or more of the following: In response to the fact that a city that failed to be attacked has been merged with other cities that were successfully attacked, the operation of merging cities that failed to be attacked and cities that were successfully attacked will not be performed this time; If a city that failed to attack already has a vassal, and the original city corresponding to the vassal is the city of the city that was successfully attacked, it is determined that there is a cyclical relationship between the failed city, the successful city, and the original city, and the operation of merging the failed city and the successful city is not performed.

5. The method according to any one of claims 2-4, characterized in that, The method further includes: Determine the number of cities in the city tree structure, and determine the number of storage areas in the warehouse; Based on the relationship between the number of cities and the number of storage areas, the city tree structure is adjusted using a corresponding preset city adjustment model to obtain the number of cities in the storage areas; this process includes: If the number of cities is greater than the number of storage areas, determine the first target city with the fewest child nodes in the city tree structure, determine other target cities whose city lords are vassals in the first target city, merge the first target city and other target cities into one city, and repeat the determination of the first target city and the merging operation until the number of cities in the storage area is obtained. If the number of cities is less than the number of storage areas, determine the second target city with the most child nodes in the city tree structure, split the vassals in the second target city to obtain multiple new cities, and repeat the determination of the second target city and the splitting operation until the number of cities in the storage area is obtained; where the city lord of the new city is the city lord of the second target city.

6. The method according to claim 5, characterized in that, The method further includes: Since there are multiple first target cities, the power value of the city lord in each first target city is determined based on the quantity of goods in each first target city, and the first target city with the smallest power value is determined from among them. Since there are multiple second target cities, the power value of the city lord in each second target city is determined based on the quantity of goods in each second target city, and the second target city with the largest power value is selected from among them.

7. The method according to claim 1, characterized in that, Based on the product clustering results and the current warehouse storage area layout information, multiple clustering layout schemes are obtained, including: The optimality of each product clustering result is determined using a pre-defined product clustering result scoring model. The optimality of each reservoir area is determined using a pre-defined reservoir area scoring model. According to the order of preference from large to small, the product clustering results and storage areas are combined in pairs to obtain a variety of clustering layout schemes.

8. The method according to claim 7, characterized in that, The step of using a preset product clustering result scoring model to determine the optimality of each product clustering result includes: For a single product clustering result, determine the number of products included in the single product clustering result; Identify the central product in the clustering results of a single product, count the number of orders that include the central product during the historical production cycle, and use the counted number of orders as the popularity of the central product; The evaluation popularity of a single product clustering result is determined based on the product outbound data of each product in the individual product clustering results. The optimality of a single product clustering result is determined based on one or more of the following: product quantity, popularity of the central product, and evaluation popularity.

9. The method according to claim 8, characterized in that, The step of determining the evaluation popularity of a single product cluster result based on the product outbound data of each product in the single product cluster result includes: Determine the average daily inventory and average daily price of each product in the historical production cycle. Based on the average daily inventory and average daily price, determine the evaluation popularity of each product. Then, accumulate them to obtain the total evaluation popularity of all products in the clustering result of a single product. Determine the popularity of the central product in the clustering results of a single product; Determine the order quantity that includes each product and every other product, and use the order quantity as the correlation between each product and every other product. Accumulate these correlations to obtain the total correlation between each product and all other products, and then accumulate them to obtain the total correlation of all products in the single product clustering result. Determine the overall association degree of the central products in the clustering results of a single product; The evaluation popularity of a single product clustering result is determined based on one or more of the following: the overall evaluation popularity of all products, the evaluation popularity of the central product, the overall correlation of all products, and the overall correlation of the central product.

10. The method according to claim 7, characterized in that, The process of using a preset reservoir scoring model to determine the optimality of each reservoir includes: For a single storage area, determine the number of available storage locations in that single storage area; Determine the distance from a single storage area to other storage areas, and then sum them up to obtain the total distance; Determine the distance from a single storage area to the target area; The preference of a single storage area is determined based on one or more of the following: the number of available storage locations, the total distance, and the distance from a single storage area to the target area.

11. The method according to claim 1, characterized in that, The step of determining the optimal clustering layout scheme from the multiple clustering layout schemes based on the model constraints includes: Extract multiple order data from the goods outbound data of historical production cycles; For a single clustering layout scheme, the storage area where the goods of each order are located is determined, so as to determine the picking distance and picking path of each order, and then the total picking path of the multiple order data is obtained; Determine the target clustering layout scheme with a closed-loop total picking path, determine the total picking distance for each target clustering layout scheme, and select the target clustering layout scheme with the smallest total picking distance as the optimal clustering layout scheme.

12. A warehouse layout planning device, characterized in that, include: The clustering module is used to obtain the goods outbound data of the warehouse in the historical production cycle, and to perform goods clustering based on the goods outbound data to obtain the goods clustering results; The scheme module is used to obtain multiple clustering layout schemes based on the product clustering results and the current storage area layout information of the warehouse; The optimization module is used to construct model constraints based on the storage area picking scenario, and determine the optimal clustering layout scheme from the multiple clustering layout schemes through the model constraints. The adjustment module is used to adjust the storage areas of goods in the current production cycle of the warehouse according to the optimal clustering layout scheme.

13. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-11.

14. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-11.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-11.

Images