Warehouse operation method and system based on storage location and picking path integrated optimization

Abstract

The application relates to the technical field of intelligent warehouse management, in particular to a warehouse operation method and system based on integrated optimization of storage positions and picking paths, which comprises the following steps: based on an initialized warehouse layout parameter set, a dynamic storage position allocation optimization model is constructed with the minimum total warehouse space waste value as an optimization target; the dynamic storage position allocation optimization model is iteratively solved to obtain an initial tray storage position allocation scheme; according to real-time order information and the initial tray storage position allocation scheme, the optimal picking path of each order is generated; according to the remaining material storage state of the warehouse after the optimal picking path is executed, the remaining material storage positions are subjected to bin secondary allocation to obtain a bin reallocation scheme, thereby generating storage position-path integrated operation instructions, and the warehouse management system operation is controlled through the storage position-path integrated operation instructions. Through the collaborative optimization of storage position allocation and picking paths, the integrated collaborative optimization of warehouse layout, storage position allocation and picking paths is realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent warehouse management technology, and in particular to a warehouse operation method and system based on integrated optimization of storage location and picking path. Background Technology

[0002] In the field of warehouse management, the rapid development of the Internet of Things (IoT), Artificial Intelligence (AI), and automation technologies is driving the innovation of warehouse management models. However, existing technologies still have some problems in warehouse management.

[0003] Traditional warehouse management often adopts a fixed storage location model, that is, each inventory unit (SKU, Stock Keeping Unit) is used. Assigning specific storage locations to units leads to low space utilization, slow response times, and difficulty in adapting to dynamic market demands. While the concept of random access storage has improved space utilization and retrieval speed, it requires a robust information system to track and manage each inventory unit in the warehouse and lacks efficient dynamic optimization methods. It has significant shortcomings, especially in the collaborative optimization of storage location allocation and picking routes. However, storage location allocation and picking route planning are often treated separately, lacking an integrated optimization model. Existing methods typically optimize storage space utilization or picking route length independently, ignoring the dynamic correlation between the two, resulting in frequent instances of idle space and overlapping routes in actual operations. Secondly, although random access storage can improve space utilization, the initially optimized storage location layout gradually becomes ineffective as inventory changes dynamically. Existing relocation strategies often focus on local adjustments and lack a global perspective for periodic optimization, resulting in cumulative efficiency losses. In addition, traditional route planning algorithms (such as the Traveling Salesman Problem (TSP) solution) often generate routes that involve complex turns and reverse movements, increasing the cognitive load on pickers and actually reducing operational efficiency.

[0004] In summary, existing technologies struggle to balance storage density and operational smoothness in warehouse management, failing to meet the real-time response requirements of large-scale warehouses in practical applications. This has become a bottleneck restricting the development of smart warehousing. Therefore, there is an urgent need for new solutions that can deeply integrate dynamic optimization of storage locations with humanized path planning to systematically improve the overall efficiency of warehousing operations. Summary of the Invention

[0005] To address the above technical problems, this invention provides a warehouse operation method and system based on integrated optimization of storage location and picking path.

[0006] In a first aspect, the present invention provides a warehouse operation method based on integrated optimization of storage location and picking path, the method comprising the following steps:

[0007] Initialize the warehouse layout based on warehouse physical structure data and pallet attribute data, and generate an initial warehouse layout parameter set;

[0008] Based on the initial warehouse layout parameter set, a dynamic storage location allocation optimization model is constructed with the goal of minimizing the total wasted space value of the entire warehouse.

[0009] The dynamic storage location allocation optimization model is iteratively solved to obtain the initial pallet storage location allocation scheme under the current warehouse layout;

[0010] In response to real-time order information, based on the real-time order information and the initial pallet storage location allocation scheme, an attention-based picking path optimization algorithm is used to generate the optimal picking path for each order.

[0011] Based on the warehouse remaining material storage status after the optimal picking route is executed, the remaining material layout data to be optimized is obtained. Based on the remaining material layout data to be optimized, the storage location of the remaining materials is re-allocated to obtain the storage location redistribution scheme.

[0012] Based on the optimal picking route and the storage location reallocation scheme, an integrated storage location-path operation instruction is generated, and the operation of the warehouse management system is controlled through the integrated storage location-path operation instruction.

[0013] In a further implementation, the step of constructing a dynamic storage location allocation optimization model based on the initialized storage layout parameter set, with the objective of minimizing the total wasted space in the entire warehouse, includes:

[0014] Define a pallet stacking identifier variable and a single pallet storage location allocation variable; wherein, the single pallet storage location allocation variable is used to indicate that a single pallet is stored alone in a storage location, and the pallet stacking identifier variable is used to indicate that two different pallets are stored in the same storage location;

[0015] Based on the initial warehouse layout parameter set and the predetermined waste value measurement rules, the total warehouse space waste value of all pallet storage location combinations is calculated using the pallet stacking identification variable and the single pallet storage location allocation variable.

[0016] A dynamic storage location allocation optimization model is constructed with the goal of minimizing the total wasted space in the entire warehouse.

[0017] In a further implementation plan, the calculation process for the total wasted space value of the entire warehouse is as follows:

[0018] When two pallets to be stored are stored together in one storage location, the space waste generated when each pallet to be stored is stacked in the same storage location is calculated based on the sum of the heights of each pallet to be stored in the same storage location, and the double pallet stacking waste value of each pallet to be stored is obtained.

[0019] When a single pallet to be stored is placed in a single pallet storage location, if it is detected that another pallet to be stored is not allowed to be placed on top of the single pallet to be stored, the restricted space waste value is obtained based on the sum of the differences between the height of all single pallet storage locations and the height of the single pallet to be stored.

[0020] Based on the pre-obtained single-pallet space waste factor, measure the single-pallet space waste value when a pallet to be stored is already stored in a location that allows two pallets to be placed, and the remaining height is insufficient to place another pallet to be stored.

[0021] The total space waste value of the entire warehouse is obtained by calculating the weighted sum of the waste value of double pallet stacking, the waste value of restricted space, and the waste value of single pallet space using the pallet stacking identification variable and the single pallet storage location allocation variable.

[0022] In a further implementation, the constraints of the dynamic storage location allocation optimization model include storage restrictions for each storage location, pallet storage feasibility constraints, space restrictions, hierarchical restrictions, and stacking rule restrictions.

[0023] The storage restriction for each storage location is specifically that the number of pallets stored in each storage location shall not exceed its maximum capacity limit;

[0024] The pallet storage feasibility constraint specifically states that each pallet can only be stored in one storage location;

[0025] The space constraint specifically states that the total height of all pallets stored in each storage location shall not exceed the preset maximum storage location height.

[0026] The hierarchical restriction constraint specifically means that the number of pallet storage layers shall not exceed its maximum allowed number of layers;

[0027] The stacking rule constraint specifically states that if neither of the two pallets is allowed to place another pallet on top of it, then the two pallets cannot be stored in the same storage location.

[0028] In a further implementation, the step of iteratively solving the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout includes:

[0029] The system scans layer by layer from the bottom to the top of the warehouse. During the scanning process, it determines the set of pallets to be stored based on the initial warehouse layout parameter set and sets the current level as the bottom layer of the warehouse.

[0030] The current level is taken as the level to be processed. The pallets of the level to be processed are selected from the set of pallets to be stored and sorted in descending order of pallet height to build a list of pallets to be processed.

[0031] Filter out all vacant storage locations in the layer to be processed and sort them in descending order of storage location height to generate a list of available locations in the layer to be processed;

[0032] Based on the list of pallets to be processed and the list of available locations, the dynamic storage location allocation optimization model is solved using a pallet location allocation algorithm to obtain a primary storage location allocation solution.

[0033] Based on a local search strategy, the pallet location reallocation solution is optimized by reallocating pallet locations to obtain the initial pallet location allocation scheme under the current warehouse layout.

[0034] In a further implementation, the step of solving the dynamic storage location allocation optimization model using a pallet location allocation algorithm based on the list of pallets to be processed and the list of available locations to obtain a primary storage location allocation solution includes:

[0035] The set of unstored pallets is initialized to an empty set, and when the list of pallets to be processed is not empty, the pallet allocation operation is performed according to the constraints of the dynamic storage location allocation optimization model, and the pallets in the list of pallets to be processed are checked in turn.

[0036] If a pallet with a height greater than the height of the first storage location in the available location list is detected in the pallet list to be processed, it is moved from the pallet list to the set of unstored pallets, and the detection of the next pallet in the pallet list to be processed continues.

[0037] Select the tallest pallet from the list of pallets to be processed as the reference pallet. Then, search for the second tallest pallet in the list of pallets to be processed. If the sum of the heights of the reference pallet and the second tallest pallet does not exceed the height of the first storage location in the list of available locations, and at least one pallet allows another pallet to be placed on top of it, then store the reference pallet and the second tallest pallet as paired pallets in the lowest available storage location; otherwise, only store the reference pallet in the lowest available storage location.

[0038] If the selected lowest available storage location is a double-depth storage location, then search the list of pallets to be processed for other stackable pallets or pallet pairs to fill the remaining space of the double-depth storage location.

[0039] Remove the selected paired pallets from the list of pallets to be processed and the set of pallets to be stored, and remove the used available storage locations from the list of available locations;

[0040] If the unstored pallet set is empty and the processing level is not the lowest level, then the pallets in the unstored pallet set are re-inserted into the processing pallet list in descending order of pallet height, the current level is adjusted to the previous level, and the available position list is regenerated.

[0041] If the set of unstored pallets is empty and the level to be processed is the lowest level, select the pallet with the smallest height from the set of unstored pallets and store it in the remaining empty storage space of the current level or the level below.

[0042] If the current level is less than the maximum level and the set of unstored pallets is empty, then update the level to be processed to the next level, and re-filter the list of pallets to be processed and available locations for the current level, and continue to allocate pallets until all pallets meet the constraints of the dynamic storage location allocation optimization model, then stop the iteration and output the initial storage location allocation solution.

[0043] In a further implementation, the step of optimizing the pallet location reallocation based on a local search strategy to obtain the initial pallet location allocation scheme under the current warehouse layout includes:

[0044] Based on the storage location type and pallet stacking status, calculate the storage location waste value of each storage location in the primary storage location allocation solution, and sort all storage locations in the primary storage location allocation solution in descending order according to the storage location waste value to generate a list of storage locations to be optimized;

[0045] Traverse the list of storage locations to be optimized, and take each storage location in the list of storage locations to be optimized as the source location, and detect feasible moves in sequence according to the four pre-built priority of move types;

[0046] If a feasible move is detected, the feasible move is executed in descending order of waste reduction magnitude, and the pallet position and storage position status are updated in real time based on the storage location information of the feasible move.

[0047] After each feasible move is executed, the total waste value of the entire warehouse space is recalculated. If the total waste value of the entire warehouse space is greater than the preset waste threshold, a new list of storage locations to be optimized is generated based on the updated storage location status. The above steps are repeated to continuously iterate and optimize until the total waste value of the entire warehouse space is less than the preset waste threshold. Then, the iterative optimization is terminated, and the initial pallet storage location allocation scheme under the current warehouse layout is output.

[0048] In a further implementation, the step of generating the optimal picking route for each order using an attention-based picking route optimization algorithm based on the real-time order information and the initial pallet storage location allocation scheme includes:

[0049] The list of pallets to be picked is obtained based on the real-time order information, and the current warehouse layout and pallet storage location information are extracted from the initial pallet storage location allocation scheme based on the list of pallets to be picked.

[0050] Based on the pallet storage location information, the current warehouse layout is converted into a graph structure to obtain a warehouse topology graph; wherein, the graph nodes of the warehouse topology graph include pallet storage location nodes, shelf aisle top and bottom nodes, and warehouse entrance and exit nodes, and the edges represent the traversable path relationships between graph nodes;

[0051] An attention mechanism is used to learn the cyclic vertical edges and cyclic horizontal edges from the warehouse topology graph, and a cyclic graph is constructed according to the order of the cyclic vertical edges and the cyclic horizontal edges. The attention mechanism includes an input layer, an embedding layer, a multi-head attention mechanism layer, a feedforward layer, a residual connection and layer normalization layer, and a linear layer.

[0052] Extract all valid picking routes from the warehouse entrance node, through each pallet storage location node, and to the warehouse exit node from the aforementioned circulation map, and select the valid picking route with the minimum total cost as the optimal picking path.

[0053] In a further implementation, the step of performing secondary allocation of storage locations for the remaining materials based on the remaining material layout data to be optimized, and obtaining a storage location reallocation scheme, includes:

[0054] Based on the remaining material layout data to be optimized, a storage and movement list is generated, and the maximum allowed execution time is defined;

[0055] The storage move operations in the storage move list are executed sequentially according to the preset move priority, and a random number is generated during the execution of each storage move. The number of storage moves equal to the number of random numbers is randomly selected from the storage move list for deletion.

[0056] During each storage move, the execution time of each storage move is calculated and accumulated. When the accumulated execution time reaches the maximum allowed execution time, the storage move is stopped, and the currently executed storage move operations are combined to form a storage move operation scheme.

[0057] Based on the total waste reduction of the storage relocation operation scheme, the optimal storage location reallocation scheme is selected from all storage relocation operation schemes.

[0058] Secondly, the present invention provides a warehouse operation system based on integrated optimization of storage location and picking path, the system comprising:

[0059] The warehouse initialization module is used to initialize the warehouse layout based on warehouse physical structure data and pallet attribute data, and generate an initial warehouse layout parameter set.

[0060] The storage location model construction module is used to construct a dynamic storage location allocation optimization model based on the initial storage layout parameter set, with the optimization objective of minimizing the total space waste value of the entire warehouse.

[0061] The storage location model solving module is used to iteratively solve the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout.

[0062] The picking route analysis module is used to respond to real-time order information and generate the optimal picking route for each order based on the real-time order information and the initial pallet storage location allocation scheme using an attention learning-based picking route optimization algorithm.

[0063] The warehouse location reallocation module is used to evaluate the remaining material layout data to be optimized based on the warehouse remaining material storage status after the optimal picking path is executed, and to perform secondary warehouse location reallocation based on the remaining material storage location to be optimized to obtain a warehouse location reallocation scheme.

[0064] The warehouse operation module is used to generate integrated storage location-path operation instructions based on the optimal picking route and the storage location reallocation scheme, and control the operation of the warehouse management system through the integrated storage location-path operation instructions.

[0065] This invention provides a warehouse operation method and system based on integrated optimization of storage location and picking path. The method initializes the warehouse layout based on warehouse physical structure data and pallet attribute data, generating an initial storage layout parameter set. Based on the initial storage layout parameter set, a dynamic storage location allocation optimization model is constructed with the goal of minimizing the total wasted space in the entire warehouse. The dynamic storage location allocation optimization model is iteratively solved to obtain an initial pallet storage location allocation scheme under the current warehouse layout. Responding to real-time order information, based on the initial pallet storage location allocation scheme, an attention-based picking path optimization algorithm is used to generate the optimal picking path for each order. Based on the remaining material storage status in the warehouse after the optimal picking path is executed, data on the remaining material layout to be optimized is obtained. The remaining material storage locations are then reassigned based on this data to obtain a storage location reallocation scheme. An integrated storage location-path operation instruction is generated based on the optimal picking path and the storage location reallocation scheme, and the operation of the warehouse management system is controlled through this integrated storage location-path operation instruction. Compared with existing technologies, this method achieves integrated and coordinated optimization of warehouse layout, storage location allocation and picking path through the coordinated optimization of storage location allocation and picking path. This effectively reduces the total space waste of the entire warehouse, improves warehouse operation efficiency and resource utilization, and thus realizes intelligent and integrated management of warehousing operations. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of a warehouse operation method based on integrated optimization of storage location and picking path provided in an embodiment of the present invention;

[0067] Figure 2 This is a schematic diagram of a shelving structure with three columns and four layers provided in an embodiment of the present invention;

[0068] Figure 3 This is a side view of a shelving structure with three columns and four layers provided in an embodiment of the present invention;

[0069] Figure 4 This is a flowchart of a warehouse operation process based on integrated optimization of storage location and picking route provided in an embodiment of the present invention;

[0070] Figure 5 This is a schematic diagram of the warehouse layout after storage location allocation provided in an embodiment of the present invention;

[0071] Figure 6 This is a schematic diagram of the picking route for the S-shaped picking strategy provided in this embodiment of the invention;

[0072] Figure 7 This is a schematic diagram of the picking route for the return picking strategy provided in an embodiment of the present invention;

[0073] Figure 8This is a schematic diagram of the maximum gap heuristic picking strategy provided in an embodiment of the present invention;

[0074] Figure 9 This is a schematic diagram of the picking route for the composite picking strategy provided in this embodiment of the invention;

[0075] Figure 10 This is a schematic diagram of the attention mechanism structure provided in an embodiment of the present invention;

[0076] Figure 11 This is a block diagram of a warehouse operation system based on integrated optimization of storage location and picking path provided in an embodiment of the present invention.

[0077] Explanation of reference numerals in the attached diagram: 101, Warehouse initialization module; 102, Storage location model construction module; 103, Storage location model solving module; 104, Picking route analysis module; 105, Warehouse location reallocation module; 106, Warehouse operation module. Detailed Implementation

[0078] The embodiments of the present invention are described in detail below with reference to the accompanying drawings. The embodiments are given for illustrative purposes only and should not be construed as limiting the present invention. The accompanying drawings are for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0079] Figure 1 This is a schematic diagram of a warehouse operation method based on integrated optimization of storage location and picking path provided by an embodiment of the present invention. The embodiment of the present invention provides a warehouse operation method based on integrated optimization of storage location and picking path, such as... Figure 1 As shown, the method includes the following steps:

[0080] S1. Initialize the warehouse layout based on the warehouse physical structure data and pallet attribute data, and generate the initial warehouse layout parameter set.

[0081] To address the storage needs of various types of materials in flexible warehouses, this embodiment utilizes standardized pallets as cargo carriers for inbound and outbound operations. All pallets have a uniform bottom dimension of 150 cm × 150 cm, but their heights vary. The warehouse's physical structure consists of multiple columns of storage locations, each configured with single-depth or double-depth shelving. The bottom dimension of each storage location is the same as the pallet's bottom dimension, with a few centimeters of operational margin reserved. For example, this operational margin can be set to 3 cm, with the specific value determined based on actual operational space requirements and ease of operation. This ensures that operators have sufficient space to work during pallet storage and retrieval. Thus, while all storage locations have the same width, their heights are differentiated based on actual storage needs and pallet height characteristics, achieving efficient utilization of storage space.

[0082] In this embodiment, storage locations are identified using column-layer two-dimensional coordinates, supporting mixed storage of multiple product categories. Furthermore, based on the depth design of the shelving, storage locations are divided into single-depth and double-depth storage locations. Single-depth storage locations are installed on single-depth shelving and can hold up to two pallets stacked vertically, defined as upper location 2 and lower location 1, respectively. Double-depth storage locations are installed on double-depth shelving, which consists of one shelving unit installed behind another without a back panel. A double-depth storage location can hold up to four pallets stacked in a front-to-back double-layer structure, designated as front lower layer location 1, front upper layer location 2, rear lower layer location 3, and rear upper layer location 4.

[0083] For shelves viewed from the front, Figure 2 This is a schematic diagram of a shelving structure with three columns and four layers provided in an embodiment of the present invention; for back-to-back double-depth shelving, Figure 3 This is a side view of a shelving structure with three columns and four layers provided in an embodiment of the present invention. The shelving has four layers, and each storage position can accommodate four pallets. Based on this, this embodiment can determine the correspondence between storage positions and shelving layout, clarify the shelving, column, and layer information to which each storage position belongs, and determine the operation rules for accessing pallets stored in the later positions 3 and 4. When the pallet stored in the earlier position 1 is not empty, the pallets stored in positions 1 and 2 must be cleared before the pallets in positions 3 and 4 can be accessed. Similarly, if a pallet is to be placed in the later position 3 or position 4, the pallets in the earlier position 1 or 2 must be cleared first.

[0084] It should be noted that single-depth racking storage is suitable for high-frequency and low-density storage scenarios. In high-frequency scenarios, for pallets containing fragile or high-value goods that cannot be stacked, the waste value is calculated by the difference between the location height and the pallet height to avoid unnecessary space occupation and ensure storage efficiency. In low-density storage scenarios, when pallets are allowed to be stacked but the remaining height is less than the minimum pallet height, the waste value is marked as 0, and the rule of not allowing stacking on top of each other is followed to meet strict constraints such as fire prevention and load-bearing capacity. Meanwhile, double-depth racking storage is suitable for high-density and mixed storage scenarios. In high-density storage scenarios, due to the larger rack depth, double-depth locations are split into independent single-depth locations to calculate waste value. For double-depth storage locations, the front depth (locations 1 and 2) and the rear depth (locations 3 and 4) are calculated separately. To maximize space utilization, the system calculates the waste value of the upper space in a dual-depth location, while adhering to the front and back pallet access order. This means that when accessing a later-depth pallet, the previous-depth pallet must be cleared first. In mixed storage scenarios, if some pallets are allowed to be stacked while others are not, when different types of pallets are stored in the same dual-depth location, the waste value of the upper space at each depth is calculated separately according to the single-pallet rule to ensure storage compliance. Furthermore, general constraint scenarios cover both layer height restrictions and compliant storage scenarios. In layer height restriction scenarios, exceeding storage limits is avoided by verifying that the product of the number of layers and the pallet height does not exceed the storage location height. In compliant storage scenarios, for pallets that are not allowed to be stacked due to batch requirements, the upper space is forcibly released through waste value calculation. That is, when such a pallet is detected, the waste value of the space above it is calculated, and it is marked as unusable to prevent unauthorized stacking and meet storage requirements.

[0085] Then, in this embodiment, the warehouse management system collects real-time physical structure data, pallet attribute data, and real-time storage status data of the warehouse, and initializes the warehouse layout based on the physical structure data, pallet attribute data, and real-time storage status data, generating an initial warehouse layout parameter set. Figure 4This is a flowchart of a warehouse operation process based on integrated optimization of storage location and picking path provided by an embodiment of the present invention. In this embodiment, the initialization of the warehouse layout parameter set includes the physical coordinates, storage capacity, and initial occupancy status of each storage location; the warehouse physical structure data includes a unique storage location number assigned to each single-depth and double-depth location, the storage location height for single-depth and double-depth storage locations, the number of storage location layers, and the shelf type used to identify the single or double depth of each storage location, etc. The unique storage location number is used to identify each storage location; for the number of storage location layers, single-depth storage locations are fixed at two layers, and double-depth storage locations are fixed at two layers each in the front and back depths; pallet attributes... The data includes pallet number, pallet height, maximum allowed stacking layers, stacking restriction flag, and minimum pallet height. The minimum pallet height is used to determine whether the remaining space can accommodate another pallet. Real-time storage status data includes real-time storage relationships and space occupancy data. The real-time storage relationships include binary variables for pallet storage location and pallet stacking relationships. The binary variable for storage location is used to track the binding relationship between pallets and storage locations. The binary variable for pallet stacking relationships is used to identify the stacking order within the same storage location. Space occupancy data includes the used height of a single-depth storage location and the used height of the front and rear depths of a double-depth storage location.

[0086] When a pallet is put on, taken off, or moved, the waste value of the corresponding storage location is recalculated. For single-depth storage locations, the waste value is calculated according to a preset waste value measurement rule formula, depending on whether one or two pallets are stored and whether stacking is allowed. For double-depth storage locations, the waste values ​​of the front and back depths are calculated separately, and the sum of the two is taken as the total waste value of the storage location. By calculating the waste value in real time, data support is provided for subsequent storage location allocation and picking route optimization.

[0087] S2. Based on the initial warehouse layout parameter set, construct a dynamic storage location allocation optimization model with the goal of minimizing the total wasted space value of the entire warehouse.

[0088] In some implementations, the step of constructing a dynamic storage location allocation optimization model based on the initial storage layout parameter set, with the objective of minimizing the total wasted space in the entire warehouse, includes:

[0089] The step of constructing a dynamic storage location allocation optimization model based on the initialized storage layout parameter set, with the objective of minimizing the total wasted space in the entire warehouse, includes:

[0090] Define a pallet stacking identifier variable and a single pallet storage location allocation variable; wherein, the single pallet storage location allocation variable is used to indicate that a single pallet is stored alone in a storage location, and the pallet stacking identifier variable is used to indicate that two different pallets are stored in the same storage location;

[0091] Based on the initial warehouse layout parameter set and the predetermined waste value measurement rules, the total warehouse space waste value of all pallet storage location combinations is calculated using the pallet stacking identification variable and the single pallet storage location allocation variable.

[0092] With the goal of minimizing the total wasted space in the entire warehouse, a dynamic storage location allocation optimization model is constructed.

[0093] In a specific embodiment, this embodiment determines the waste value measurement rules based on the storage location type and pallet stacking status. The waste value measurement rules are as follows:

[0094] If the storage space is empty, the waste value is zero;

[0095] If a storage location has only one pallet and allows another pallet to be placed on top of it, the waste value is zero when the difference between the storage location height and the storage pallet height is greater than or equal to the height of the lowest pallet; otherwise, the waste value is the storage location height minus the pallet height.

[0096] If a single pallet i is stored separately in the storage location Furthermore, if it is not allowed to place another pallet on top of it, the waste value is the storage space height minus the pallet height;

[0097] If two pallets i and pallet k are stacked in the storage location In this context, the waste value is the storage location height minus the sum of the heights of the two pallets.

[0098] This embodiment treats each dual-depth storage location as two independent single-depth storage locations to simplify the model complexity, while preserving the access order constraints between the front and rear depths. For any pallet and any location, this embodiment pre-calculates the single-pallet space waste factor. Single pallet space waste coefficient This is used to measure waste when one pallet is already stored in a location that allows two pallets and there is not enough space to place another pallet. The single-pallet space waste factor is calculated if placing another pallet on top of the pallet does not allow placing another pallet on top of it, or if the pallet height plus the minimum pallet height is less than or equal to the storage location height. The value is zero; otherwise, the single-pallet space waste factor is zero. The space wastage factor per pallet is calculated by subtracting the pallet height from the storage height. The specific calculation is as follows: If =1 or +minc≤ ,but =0; otherwise = ,in, The stacking limit identifier for pallet i is minc, which is the minimum pallet height. It represents the minimum value of all pallet heights and is used to determine whether there is enough space left to store another pallet. The height of tray i For storage The height.

[0099] This embodiment calculates the storage combination of all pallets to be stored in the warehouse based on the initialized warehouse layout parameter set and the pre-determined waste value measurement rules. For the case where two pallets to be stored are stored in the same storage location, this embodiment calculates the space waste generated when each pallet to be stored is stacked in the same storage location based on the sum of the heights of each pallet to be stored in the same storage location. Specifically, for any two pallets to be stored i and pallets to be stored k ( When pallet i and pallet k are stored together in the storage location At that time, the space wastage caused by stacking each pallet to be stored is valued at ,in, For storage height, The height of the tray i to be stored is... Given the height of the pallet to be stored k, this embodiment obtains the stacking waste value of each double pallet for each pallet to be stored by traversing all possible pallet combinations. When a single pallet to be stored is placed in a single pallet storage location, if it is detected that the pallet does not allow another pallet to be stored to be placed on top of it, the sum of the differences between the height of the single pallet storage location and the height of the single pallet to be stored is calculated to obtain the restrictive space waste value. At the same time, this embodiment measures the single pallet space waste value when a pallet to be stored has been stored in a location that allows the placement of two pallets and no other pallet to be stored can be placed, based on the pre-obtained single pallet space waste coefficient.

[0100] Then, in this embodiment, a tray stacking identifier variable is defined. Pallet storage location allocation variables All variables are binary variables. Among them, the pallet stacking identifier variable is used to indicate that two different pallets are stored in the same storage location. If pallet i and pallet k are stacked in the storage location... Then the tray stacking identifier variable Set to 1, otherwise the tray stacking indicator variable. Set to 0; the pallet storage location allocation variable is used to represent a single pallet stored in a storage location. If pallet i is stored in storage location... Then the pallet storage location allocation variable Set to 1, otherwise the pallet storage location allocation variable Take 0.

[0101] Next, this embodiment utilizes the tray stacking identifier variable. Pallet storage location allocation variables The space waste values ​​of the above three parts are weighted and summed to obtain the total space waste value of the entire warehouse. A dynamic storage location allocation optimization model is constructed with the goal of minimizing the total space waste value. In this embodiment, the dynamic storage location allocation optimization model is modeled as an integer linear programming problem, and its objective function is as follows:

[0102]

[0103] In the formula, For storage The height indicates the storage level. Vertical space capacity; The height of tray i represents the vertical space occupied by tray i. The height of pallet k represents the vertical space occupied by pallet k. This is a binary variable indicating the stacking of pallets. If pallet i and pallet k are stacked in storage location... In the middle, then ,otherwise ; This is a Boolean variable that indicates the stacking limit for tray i. This indicates that pallet i is not allowed to be stacked on top of other pallets. This indicates that stacking is permitted; Assign a variable to the pallet storage location; this variable is a binary variable. If pallet i is stored alone in storage location... In the middle, then ,otherwise ; The single-pallet space waste factor represents the space wasted in the storage location. The waste value generated when pallet i is stored alone; m is the total number of storage locations, i.e., the total number of storage locations in the warehouse that can be used to store pallets; n is the total number of pallets; j is the storage location index; i is the pallet index; k is the pallet index, which usually represents another pallet stacked with pallet i, where... .

[0104] The objective function of the dynamic storage location allocation optimization model represents minimizing the total space waste value of the entire warehouse. The first term is the sum of the waste values ​​of all double-pallet stacking, the second term is the sum of the waste values ​​of all restricted space, and the third term is the sum of the waste values ​​of all single-pallet space. The constraints of the dynamic storage location allocation optimization model include storage restrictions for each storage location, pallet storage feasibility constraints, space restrictions, hierarchical restrictions, and stacking rule restrictions. The mathematical expression for the storage restriction constraint of each storage location is:

[0105]

[0106] In this embodiment, the storage restriction constraint for each storage location is that the number of pallets stored in each storage location does not exceed the maximum capacity limit of its storage location, ensuring that each storage location can store at most one pallet or a pair of stacked pallets. Specifically, a single-depth storage location can store at most two pallets, and the front and rear depths of a double-depth storage location are independently restricted according to the single-depth storage location.

[0107] The mathematical expression for the feasibility constraint of pallet storage is:

[0108]

[0109] The pallet storage feasibility constraint specifically states that each pallet is stored in only one storage location, meaning that each pallet must be assigned to one and only one storage location to ensure that all pallets are stored properly.

[0110] The mathematical expression for spatial constraints is:

[0111]

[0112] The space constraint specifically states that the total height of all pallets stored in each storage location does not exceed the preset maximum storage location height, ensuring that the pallets are not stored beyond the storage space.

[0113] The mathematical expression for hierarchical constraints is:

[0114]

[0115] The hierarchical restriction specifically stipulates that the number of pallet storage layers shall not exceed its maximum allowed number of layers. Follow the pallet storage hierarchy requirements.

[0116] The mathematical expression for the stacking rule constraint is:

[0117]

[0118] The stacking rule constraint specifically states that if neither of the two pallets is allowed to place another pallet on top of it, then the two pallets cannot be stored in the same storage location, ensuring that the pallet stacking complies with the rule.

[0119] The mathematical expression for the variable definition constraint is:

[0120]

[0121]

[0122] In the formula, For storage The number of layers, of which single-depth storage is fixed at 2 layers, and double-depth storage has 2 layers before and after, that is, a total of 4 layers for double-depth storage. The maximum number of layers that pallet i is allowed to be stacked on represents the highest level that pallet can be stacked on top of other pallets in the storage location; This is a stacking restriction indicator for tray k.

[0123] S3. Iteratively solve the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout.

[0124] In some implementations, the step of iteratively solving the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout includes:

[0125] The system scans layer by layer from the bottom to the top of the warehouse. During the scanning process, it determines the set of pallets to be stored based on the initial warehouse layout parameter set and sets the current level as the bottom layer of the warehouse.

[0126] The current level is taken as the level to be processed. The pallets of the level to be processed are selected from the set of pallets to be stored and sorted in descending order of pallet height to build a list of pallets to be processed.

[0127] Filter out all vacant storage locations in the layer to be processed and sort them in descending order of storage location height to generate a list of available locations in the layer to be processed;

[0128] Based on the list of pallets to be processed and the list of available locations, the dynamic storage location allocation optimization model is solved using a pallet location allocation algorithm to obtain a primary storage location allocation solution.

[0129] Based on a local search strategy, the pallet location reallocation solution is optimized by reallocating pallet locations to obtain the initial pallet location allocation scheme under the current warehouse layout.

[0130] In some implementations, the step of solving the dynamic storage location allocation optimization model using a pallet location allocation algorithm based on the list of pallets to be processed and the list of available locations to obtain a primary storage location allocation solution includes:

[0131] The set of unstored pallets is initialized to an empty set, and when the list of pallets to be processed is not empty, the pallet allocation operation is performed according to the constraints of the dynamic storage location allocation optimization model, and the pallets in the list of pallets to be processed are checked in turn.

[0132] If a pallet with a height greater than the height of the first storage location in the available location list is detected in the pallet list to be processed, it is moved from the pallet list to the set of unstored pallets, and the detection of the next pallet in the pallet list to be processed continues.

[0133] Select the tallest pallet from the list of pallets to be processed as the reference pallet. Then, search for the second tallest pallet in the list of pallets to be processed. If the sum of the heights of the reference pallet and the second tallest pallet does not exceed the height of the first storage location in the list of available locations, and at least one pallet allows another pallet to be placed on top of it, then store the reference pallet and the second tallest pallet as paired pallets in the lowest available storage location; otherwise, only store the reference pallet in the lowest available storage location.

[0134] If the selected lowest available storage location is a double-depth storage location, then search the list of pallets to be processed for other stackable pallets or pallet pairs to fill the remaining space of the double-depth storage location.

[0135] Remove the selected paired pallets from the list of pallets to be processed and the set of pallets to be stored, and remove the used available storage locations from the list of available locations;

[0136] If the unstored pallet set is empty and the processing level is not the lowest level, then the pallets in the unstored pallet set are re-inserted into the processing pallet list in descending order of pallet height, the current level is adjusted to the previous level, and the available position list is regenerated.

[0137] If the set of unstored pallets is empty and the level to be processed is the lowest level, select the pallet with the smallest height from the set of unstored pallets and store it in the remaining empty storage space of the current level or the level below.

[0138] If the current level is less than the maximum level and the set of unstored pallets is empty, then update the level to be processed to the next level, and re-filter the list of pallets to be processed and available locations for the current level, and continue to allocate pallets until all pallets meet the constraints of the dynamic storage location allocation optimization model, then stop the iteration and output the initial storage location allocation solution.

[0139] The pallet location allocation algorithm in this embodiment constructs an initial solution to reserve lower-level space for pallets with hierarchical restrictions. To address this issue, the construction process employs a bottom-to-top layer-by-layer traversal strategy. For each given level, this embodiment prioritizes arranging the necessary pallets at that level. Typically, pallets without hierarchical restrictions are placed at the highest level. To minimize space waste during the hierarchical analysis, this embodiment considers only empty positions. Specifically, empty positions in the current level are sorted from highest to lowest height, and pallets are also sorted according to this standard. When processing pallet storage, the first pallet in the sorted pallet list is selected first. The pallet location allocation algorithm then searches for another pallet in this list, which must meet two key conditions: first, the sum of the heights of the two pallets does not exceed the first empty position in the list. The algorithm considers two key conditions: first, the height of the pallet; second, at least one pallet must allow another pallet to be stacked on top of it; if a pallet that meets both conditions is found, the two pallets are placed in the same location; if no such pallet combination exists, only the first pallet considered is placed. When placing pallets, the algorithm selects the lowest-height location from the sorted list that can accommodate the pair of pallets or a single pallet. If the selected location is a double-depth location, the pallet placement algorithm will further attempt to find another pallet or two pallets in the list to achieve complete occupancy of the double-depth location. If a pallet cannot be placed in the current layer, the algorithm will transfer its storage requirement to the next layer for further attempts. During the entire algorithm execution, the Notstored set is used to record the pallets that have not yet been placed in the current solution, and this set will change dynamically as the algorithm progresses; while the Incoming Pallets set remains unchanged, always representing the set of incoming pallets to be processed.

[0140] In this specific embodiment, the first step is to perform an initialization operation. The Notstored set of pallets to be stored is initialized as the Incoming Pallets set, and the current level is set to the bottom layer of the warehouse (current=1). The level variable h is set to the current level. Next, this embodiment constructs a PalletsList to be processed, which contains pallets in the Notstored set with a maximum storage level of h, and the pallets in the list are arranged in descending order of height. At the same time, the Imp set of unstored pallets is initialized to an empty set. Then, this embodiment constructs a LocationsList to be used, which contains all empty locations at level h, and the locations in the list are arranged in descending order of height. After that, the pallet list is processed in a loop. When the PalletsList to be processed is not empty, the following operations are performed:

[0141] If a pallet in the PalletsList has a height greater than the height of the first location in the LocationsList, move that pallet from the PalletsList to the Imp set of unstored pallets, and select the first pallet in the PalletsList. If a next pallet exists in the PalletsList such that the sum of the heights of the two pallets does not exceed the height of the first location in the LocationsList, and at least one pallet allows another pallet to be stacked on top of it, select that next pallet. Remove the selected pallets from the PalletsList and the Notstored set of unstored pallets. Store the selected pallets (or pallet pairs) in an empty location in the LocationsList that is taller than the selected location. If the selected location supports four pallets, select a suitable pallet or pallet pair from the PalletsList and place it in that location to minimize space waste, and remove these selected pallets from the PalletsList and the Notstored set of unstored pallets. In `ist`, the positions of already stored pallets are removed. Then, unstored pallets are processed. If the unstored pallet set `Imp` is empty and the current processing level `h` is greater than 1, the pallets from `Imp` are inserted into the appropriate positions in the pallets list `PalletsList`, and the level variable `h` is decremented by 1 to allow processing to continue at a higher level. The list of available positions is then reconstructed. Finally, the last pallet is processed. If the unstored pallet set `Imp` is empty and the current processing level has reached the lowest level (h=1), a pallet is selected from the unstored pallet set `Imp` and attempted to be stored in some empty positions at the current level or lower to minimize space waste. The level variable `h` is then set to... The value of (current+1) is used to move the level down for further processing, and the current variable is incremented by 1. If the level variable h is less than the maximum number of levels, the pallet list is reconstructed. When all pallets meet the constraints of the dynamic storage location allocation optimization model (including height restrictions, stacking rules, etc.), the iteration is terminated. Through the above steps, the initial storage location allocation solution is finally obtained. The initial storage location allocation solution includes, but is not limited to, the initial feasible solution of pallet position and stacking relationship. This embodiment effectively generates an initial storage location allocation scheme that meets the goal of minimizing space waste under the premise of ensuring storage compliance through a systematic level scanning and dynamic allocation mechanism, providing a high-quality initial solution for subsequent optimization.

[0142] In some implementations, the step of optimizing the pallet location reallocation based on a local search strategy to obtain an initial pallet location allocation scheme under the current warehouse layout includes:

[0143] Based on the storage location type and pallet stacking status, calculate the storage location waste value of each storage location in the primary storage location allocation solution, and sort all storage locations in the primary storage location allocation solution in descending order according to the storage location waste value to generate a list of storage locations to be optimized;

[0144] Traverse the list of storage locations to be optimized, and take each storage location in the list of storage locations to be optimized as the source location, and detect feasible moves in sequence according to the four pre-built priority of move types;

[0145] If a feasible move is detected, the feasible move is executed in descending order of waste reduction magnitude, and the pallet position and storage position status are updated in real time based on the storage location information of the feasible move.

[0146] After each feasible move is executed, the total waste value of the entire warehouse space is recalculated. If the total waste value of the entire warehouse space is greater than the preset waste threshold, a new list of storage locations to be optimized is generated based on the updated storage location status. The above steps are repeated to continuously iterate and optimize until the total waste value of the entire warehouse space is less than the preset waste threshold. Then, the iterative optimization is terminated, and the initial pallet storage location allocation scheme under the current warehouse layout is output.

[0147] In a specific embodiment, the pallet location allocation algorithm mainly optimizes the storage of pallets with hierarchical constraints. However, in this embodiment, when using the pallet location allocation algorithm, the lower-level locations may not be fully utilized in the primary storage location allocation solution, leading to space waste. To solve this problem, this embodiment uses a local search algorithm to continuously execute a set of movement operations selected from four specific types to reduce space waste. Each type must satisfy the target storage location capacity limit and pallet storage constraints. Three of these four movement types involve operations on two pallets, and the remaining one involves operations on a single pallet. During the optimization process, all double-depth locations are treated as two independent single-depth locations. Specifically, these four movement types are:

[0148] Movement Type 1: Move two trays that are stacked at one source location to the same target location;

[0149] Movement type 2: Transferring a pallet from one source location to another destination location;

[0150] Movement type 3: Move two pallets located in different positions to a new empty target location and stack one pallet on top of the other;

[0151] Movement Type 4: Exchange a pallet that is stored alone at the source location and on which no pallet can be placed, with two pallets at the destination location that can be placed on which no pallet can be placed.

[0152] When performing these movement operations, this embodiment only considers feasible movements that can change the position of pallets in the IncomingPallets set. That is, the target location must have sufficient capacity to store the pallets and meet relevant pallet constraints. Furthermore, these movement operations are only performed if they can reduce waste. The fourth movement type also accepts movements that do not change the available centimeters, as such movements may enable other types of improved movements in the future. In this embodiment, the local search algorithm first constructs a waste list for the source location and sorts it from highest to lowest waste level. If there are identical waste values, they are handled randomly. Subsequently, the local search algorithm... The search algorithm iterates through the list four times, searching for one type of move each time. When an improved move is identified at a source location, further investigation of that location is stopped, and the move operation is stored. The location information involved in the move is marked and will not be investigated again. The search for the target location is always conducted in a list of possible target locations sorted by height in ascending order (this list includes empty locations, fully occupied locations, or partially occupied locations). If the same situation exists, it is resolved randomly. If an improved move is stored, these moves are executed and the search process restarts. For each type of move, the list of possible target locations consists of unmarked locations that meet the move requirements. Specifically:

[0153] For Type 1 movement, the source position is selected starting from the first position in the list of possible source positions. After the source position is selected, the target position is selected from the list of possible target positions, starting from the first position whose height is less than or equal to the height of the source position.

[0154] For type 2 moves, the source location is also selected from the first position in the list of possible source locations. If the selected source location has only one tray, an empty target location is searched. If no improvement move is found, the program continues to search for a location containing a single tray. If the source location contains two trays and no improvement move is found, the same procedure is repeated for the other tray.

[0155] Type 3 movement is performed by matching the position in each of the possible source locations list with the next position, during which at least one pallet must allow another pallet to be placed on top of it, and the target location is selected from the first of the possible target locations list that can accommodate two pallets;

[0156] For type 4 movement, the source location is selected from the first possible source location in the list that contains a single pallet that does not allow a pallet to be placed on it, and the target location is selected from the first possible target location in the list that contains two pallet locations that both allow a pallet to be placed on them.

[0157] If the total improvement from moving stores during the entire local search process is equal to the current total waste, then the store move is executed all at once, and the local search stops immediately, because the waste will obviously be reduced to zero at this point.

[0158] The specific steps in optimizing the pallet location reallocation based on the primary storage location allocation solution using a local search strategy to obtain the initial pallet storage location allocation scheme under the current warehouse layout are as follows:

[0159] This embodiment first performs a location list construction operation, initializing an empty list LocList as the list of storage locations to be optimized. Then, it iterates through all locations in the warehouse layout, and for each storage location... The primary storage location allocation solution is calculated for each storage location based on its type (single-depth or double-depth) and pallet stacking status. The resulting storage space waste value ,like If the waste value is found, it indicates that there are wasted storage locations. All wasteful storage locations and their corresponding waste values ​​are added to the LocList of storage locations to be optimized. After traversing all positions, the LocList is sorted from highest to lowest waste value. Then, this embodiment uses four priority-based move operation types to optimize the storage locations. The LocList is traversed sequentially, with each unmarked storage location (unmarkedele) used as the source position. The initial value range of q is set to... Then, a loop process is entered. For each movement type q, feasible optimized movements are searched in the unmarked target locations. In this loop process, this embodiment analyzes whether there is an improved movement with unmarkedelem as the source location and another unmarked location as the target location. If such an improved movement exists, the found movement operation is stored in the movement list StoredMovements, and the location information involving the movement operation is marked. At the same time, it is checked whether the sum of the improvements of the stored movement operations is the same as the total waste value of the current warehouse layout. If they are the same, the subsequent steps of executing the stored movement operation are entered. If there is no improved movement, unmarkedelem is updated to the next unmarked element in the list LocList, and the research operation continues until all unmarked elements are traversed.

[0160] Finally, this embodiment performs storage movement operations. During the execution phase, all movements in the movement list StoredMovements are traversed in descending order of waste reduction. For each movement, a corresponding pallet location movement operation is performed, and the pallet location and storage location status are updated in real time. At the same time, the results of movement execution are recorded according to actual needs. After each round of movement, the total space waste value of the entire warehouse is recalculated. If the total waste value still exceeds the preset threshold, the storage location list LocList to be optimized is regenerated based on the latest storage location status, and the above optimization process is repeated. Iteration continues until the total waste value is lower than the threshold. After all storage movement operations are completed, the movement list StoredMovements is cleared to prepare for subsequent optimization operations. If any movement operation results in a waste value of the warehouse layout greater than 0, the location list construction operation step is returned, and the optimization process is repeated. If the waste value is 0 or the preset optimization target is reached after all movement operations are executed, the entire pallet location reallocation optimization process ends. The pallet storage location allocation scheme obtained at this time is the initial pallet storage location allocation scheme under the current warehouse layout.

[0161] S4. In response to real-time order information, based on the real-time order information and the initial pallet storage location allocation scheme, an attention-based picking path optimization algorithm is used to generate the optimal picking path for each order.

[0162] In some implementations, the step of generating the optimal picking route for each order using an attention-based picking route optimization algorithm based on the real-time order information and the initial pallet storage location allocation scheme includes:

[0163] The list of pallets to be picked is obtained based on the real-time order information, and the current warehouse layout and pallet storage location information are extracted from the initial pallet storage location allocation scheme based on the list of pallets to be picked.

[0164] Based on the pallet storage location information, the current warehouse layout is converted into a graph structure to obtain a warehouse topology graph; wherein, the graph nodes of the warehouse topology graph include pallet storage location nodes, shelf aisle top and bottom nodes, and warehouse entrance and exit nodes, and the edges represent the traversable path relationships between graph nodes;

[0165] An attention mechanism is used to learn the cyclic vertical edges and cyclic horizontal edges from the warehouse topology graph, and a cyclic graph is constructed according to the order of the cyclic vertical edges and cyclic horizontal edges. The attention mechanism includes an input layer, an embedding layer, a multi-head attention mechanism layer, a feedforward layer, a residual connection and layer normalization layer, and a linear layer.

[0166] Extract all valid picking routes from the warehouse entrance node, through each pallet storage location node, and to the warehouse exit node from the aforementioned circulation map, and select the valid picking route with the minimum total cost as the optimal picking path.

[0167] In this specific embodiment, after completing the storage location allocation, it is necessary to optimize the processing path after the order arrives. Figure 5 This is a schematic diagram of the warehouse layout after storage location allocation provided in this embodiment of the invention. Based on the storage location allocation, the current warehouse layout consists of multiple parallel vertical aisles. Items are stored on both sides of the aisles, and there is a horizontal aisle (one up and one down) at the top and bottom of each aisle to facilitate navigation between the aisles for picking personnel. A single warehouse is also provided. Assuming the aisle width is negligible, the horizontal distance within the aisles does not need to be considered. Based on this, this embodiment first constructs a graphical representation after storage location allocation. It should be noted that the single picker path problem can be considered a variant of the traveling salesman problem, where items and warehouses are treated as nodes. An effective picking route starts from the warehouse, passes through each item exactly once, and returns to the warehouse. The single picker path problem can also be represented as the Steiner traveling salesman problem, where items are nodes that can be visited multiple times, while the top and bottom of the aisles are nodes that do not need to be visited. However, to maintain consistency with existing deep reinforcement learning research results, the standard TSP model is used here for explanation. The four common heuristic picking route patterns are S-shaped, return, midpoint, maximum gap, and composite picking strategies. The following will briefly describe the four heuristic picking route patterns:

[0168] S-shaped picking strategy: The S-shaped heuristic algorithm, also known as the traversal method, is one of the simpler picking strategies. It requires the picker to enter each aisle containing at least one item to be picked, completely traverse the aisle, and leave from the other end. This process is repeated for all aisles containing items before returning to the warehouse. Figure 6 This is a schematic diagram of the picking route for the S-shaped picking strategy provided in this embodiment of the invention;

[0169] Back-to-back picking strategy: For each aisle containing the required items, the picker enters and exits from the front across the aisle, collecting all items on the picking list. Figure 7 This is a schematic diagram of the picking route for the return picking strategy provided in an embodiment of the present invention;

[0170] The maximum clearance heuristic strategy involves the picker completely traversing the first and last aisles containing items. For each intermediate aisle, the picker enters and exits from the bottom and top in a specific manner, maximizing the distance of untraversed aisles while ensuring access to all items. Figure 8 This is a schematic diagram of the maximum gap heuristic picking strategy provided in an embodiment of the present invention;

[0171] Composite picking strategy: Combining the advantages of return picking and S-shaped picking strategies, for all adjacent aisles containing items in the picking list, a heuristic algorithm is selected to minimize the movement distance between the furthest picks. Figure 9 This is a schematic diagram of the picking route for the composite picking strategy provided in this embodiment of the invention;

[0172] Optimal Picking Strategy: An optimal procedure for picker paths in rectangular warehouses, which can be linearly solved relative to the number of aisles and picking locations. The algorithm sequentially constructs a partially circulating subgraph, where nodes consist of the top and bottom of each aisle and the location of each item to be picked.

[0173] This embodiment extracts a list of pallets to be picked based on real-time order information. The real-time order information includes the type and quantity of items to be picked, as well as the corresponding storage location. The corresponding storage location coordinates are obtained from the initial storage location allocation scheme. The current warehouse layout consists of multiple parallel and vertical aisles, with storage locations set on both sides of the aisles. A horizontal connecting passage is set at the top and bottom of each aisle to facilitate navigation between aisles for pickers. Assuming that the width of the aisles is negligible, the horizontal distance within the aisles does not need to be considered. This embodiment abstracts the physical layout of the warehouse into a warehouse topology graph structure. The graph nodes include pallet storage location nodes, top and bottom nodes of shelf aisles, and warehouse entrance and exit nodes. Edges represent the traversable path relationships between nodes, and distance weights reflect the actual movement costs. This embodiment transforms the single picker path problem into an improved variant of the traveling salesman problem. An effective picking route starts from the warehouse, passes through each item exactly once, and returns to the warehouse. This embodiment treats the item to be picked and the warehouse as mandatory nodes, and adds the end of the aisle as an optional access point. Starting from the warehouse, traversing all item points and returning, the total path is the shortest.

[0174] This embodiment utilizes an attention mechanism to approximate a strategy, sequentially outputting the vertical and horizontal edges to be added to the cycle graph until a complete cycle graph is constructed. The nodes of the cycle graph consist of the top and bottom nodes of the shelf aisles and each pallet storage location node to be picked. In this embodiment... Figure 10This is a schematic diagram of the attention mechanism structure provided in an embodiment of the present invention. The structure of the attention mechanism mainly includes an input layer, an embedding layer (corresponding to input embedding), channel encoding, a multi-head attention mechanism, a feedforward layer, residual connections, layer normalization, and a linear layer. The input signal is first encoded and converted by the embedding layer, and then fused with the result of the channel encoding (represented by ⊕ in the figure). Then it enters the multi-head attention mechanism module, so that the multi-head attention mechanism can extract features and model relationships of the input signal, capturing the dependencies between different nodes. The output of the multi-head attention mechanism passes through the feedforward layer, residual connections, and layer normalization in sequence. The process involves performing residual connections and layer normalization operations once between the multi-head attention mechanism and the feedforward layer, and once between the feedforward layer and subsequent layers. Finally, the linear layer maps the final features to the output space, generates an edge addition probability distribution, and dynamically selects the optimal circulating vertical edge (movement within the passageway) and circulating horizontal edge (transfer between passageways) based on the edge addition probability distribution. This process gradually expands the partially circulating subgraph until it completely covers all nodes. All valid path schemes are extracted from the generated circulating graph. In this embodiment, the total cost of each valid path scheme is calculated. The total cost is the sum of the base cost and the penalty term, where the base cost is the sum of the edge weights (reflecting the travel distance). The penalty term is the product of the number of turns and the preset weight coefficient, which is dynamically adjusted by the attention weight entropy value. In this embodiment, the efficient path scheme with the minimum total cost is selected as the final picking path. In this embodiment, the embedding layer converts the input data into a vector representation suitable for model processing; the channel encoding provides the model with prior information such as warehouse aisle layout; the multi-head attention mechanism can capture the complex dependencies between different positions in the input sequence, enabling the model to better understand the relationship between warehouse layout and item location; the feedforward layer performs nonlinear transformation on the output of the attention mechanism to enhance the model's expressive power; residual connections and layer normalization help alleviate the gradient vanishing and gradient exploding problems, accelerate model training, and improve model stability; the linear layer converts the model's output into the final decision result. This method determines which edges to add to the circuit graph in sequence, providing a different perspective for solving the problem. Compared with the traditional method of directly outputting node arrangement, this embodiment focuses on the key relationships between different picking points through the attention mechanism, gradually deciding to add vertical or horizontal connection edges, and dynamically generating efficient picking paths. It is especially suitable for scenarios such as the single picker path problem, where the connections between nodes are not limited to simple arrangement relationships, but also need to consider practical factors such as warehouse layout.

[0175] S5. Based on the warehouse remaining material storage status after the optimal picking route is executed, evaluate the remaining material layout data to be optimized, and perform secondary allocation of the remaining material storage locations based on the remaining material layout data to be optimized, to obtain a warehouse location redistribution scheme.

[0176] In some implementations, the step of performing secondary allocation of storage locations for remaining materials based on the remaining material layout data to obtain a storage location reallocation scheme includes:

[0177] Based on the remaining material layout data to be optimized, a storage and movement list is generated, and the maximum allowed execution time is defined;

[0178] The storage move operations in the storage move list are executed sequentially according to the preset move priority, and a random number is generated during the execution of each storage move. The number of storage moves equal to the number of random numbers is randomly selected from the storage move list for deletion.

[0179] During each storage move, the execution time of each storage move is calculated and accumulated. When the accumulated execution time reaches the maximum allowed execution time, the storage move is stopped, and the currently executed storage move operations are combined to form a storage move operation scheme.

[0180] Based on the total waste reduction of the storage relocation operation scheme, the optimal storage location reallocation scheme is selected from all storage relocation operation schemes.

[0181] The SLAP (Packet Location Allocation) algorithm can also be used to determine better storage locations for existing pallets. Specifically, it assumes the warehouse is empty and reallocates all existing pallets. At this point, the Incoming Pallets set equals the stored pallets. By comparing the difference in space availability metrics before and after the process, the achievable improvement gap is determined. Managers can then decide, based on the importance of this gap, whether to reduce it by relocating some pallets during periods of low workload. Typically, it's impossible to relocate all pallets; therefore, managers may focus on the following three issues: first, minimizing the number of relocations, i.e., performing a fixed number of relocations to reduce current space waste; second, maximizing waste reduction with a fixed number of relocations, i.e., obtaining the maximum waste reduction after performing a fixed number of relocations; and third, maximizing waste reduction within a given timeframe, i.e., achieving the maximum waste reduction effect within a certain time period.

[0182] To address the three scenarios described above, this embodiment employs a local search algorithm to use the current pallet position as the initial solution. It's important to note that the same item is not moved multiple times. If the goal is to minimize the number of moves, storage moves are executed in order of decreasing improvement until the desired level of improvement is achieved or all moves have been completed. If the goal is to maximize waste reduction with a fixed number of moves, only the optimal storage move is performed. Furthermore, all pallets other than the optimal fixed move pallet can be used. These algorithms may place the pallet in a new location. The difference between the waste of the current solution and the optimal solution indicates the potential improvement after a maximum of several moves. Different rules for selecting fixed pallets may produce different solutions; for example, pallets located in locations with less waste or those with strict hierarchical restrictions can be selected.

[0183] If the objective is to maximize waste reduction within a given time, the execution time of relocation needs to be determined. The execution time of a relocation is the sum of the fulfillment time between two locations, the input / output time, and the transfer time. The transfer time is the time required to load and unload pallets involved in the relocation. However, the travel time and input / output time between locations are negligible compared to the time required for a forklift to store or retrieve a unit load at a storage location. Therefore, this process only considers the transfer time. One solution is to consider storage moves using a local search algorithm and allocate them to forklifts, ensuring that the execution time allocated to any vehicle does not exceed a set maximum time limit. To this end, this embodiment proposes three deterministic allocation rules: prioritizing the move with the greatest improvement; prioritizing the move with the shortest execution time; and prioritizing the move with the highest improvement-to-execution-time ratio.

[0184] To evaluate performance, the results generated according to these rules are compared with the sampling algorithm. A fixed number of solutions are obtained iteratively by performing local search on the moves in the stored move list. The specific steps are as follows: Define nsolution as the required number of solutions and pre-set a rule to sort the stored move list according to the rule. In the iterative solution generation phase, for each iteration, moves are executed sequentially from the sorted move list, ensuring that the maximum time limit is not exceeded. Depending on the actual situation, some moves may be deleted from the list, while other moves are reintroduced into the list. After each successful generation of a new solution, the performance index of the solution is recorded. Finally, the solutions obtained by the sampling algorithm and their performance evaluation results are output.

[0185] In a specific embodiment, during the secondary allocation of storage locations for remaining materials based on the remaining material layout data to be optimized, this embodiment generates a storage movement list based on the remaining material layout data to be optimized. This storage movement list contains all possible storage movement operations. Considering the generated storage movement list, the total number of movement operations is counted and marked as nstored. Simultaneously, for each storage movement operation in the storage movement list... In this embodiment, v( ) is assigned to the first according to the preset sorting rules. Given a priority value for each move operation, sort the moves according to priority to form an ordered move list L={ , ,..., ,..., } where nstored represents the total number of move operations, and the maximum allowed execution time for the entire storage location redistribution process is set to T. The maximum allowed execution time is used to limit the execution duration of the entire move operation. Then, in this embodiment, the following operation is repeated nsolution times to generate multiple storage location redistribution schemes. In the initialization phase of each repetition, for q values ​​from 1 to 4, unmarkedelem is set to the first unmarked element in the storage location list LocList to be optimized, ensuring that there is a marked state at the beginning of each iteration. Then, the inner repetition step is entered, and the storage move operations are executed sequentially according to the order of the ordered move list L until the cumulative execution time reaches the maximum allowed execution time T. During this process, regardless of whether the move has been actually executed, each checked move needs to be marked for subsequent processing.

[0186] Meanwhile, in this embodiment, a random number num is generated in each iteration. Based on the generated random number num, an equal number of storage moves are randomly selected from the storage moves checked in the last four iterations. These selected storage moves are temporarily deleted from the ordered move list L. These deleted storage moves will be reintroduced into the ordered move list L after six iterations to ensure the diversity of solutions. During each iteration of generating a solution, the total waste reduction of the solution is recorded. Finally, from all generated storage move operation solutions, the optimal warehouse location reallocation solution is selected based on the key indicator of total waste reduction. This method can fully consider the time constraints and efficiency requirements of warehouse operations. Through this structured secondary allocation strategy, the optimal adjustment of the storage location of remaining materials can be achieved within a limited time, effectively improving warehouse space utilization and reducing operating costs. It is suitable for the dynamic optimization needs of large-scale warehousing systems.

[0187] S6. Generate integrated storage location-path operation instructions based on the optimal picking route and the storage location reallocation scheme, and control the operation of the warehouse management system through the integrated storage location-path operation instructions.

[0188] The Storage Location Allocation Problem (SLAP) aims to maximize available storage space in chaotic warehouse systems. These systems are characterized by dynamic storage, employ a single-command model, use pallets as storage units, and rely on forklifts for access. However, the SLAP problem focuses solely on maximizing available storage space. To address this, this embodiment proposes a heuristic algorithm and a local search algorithm. Furthermore, it couples an attention-based picking path optimization method, achieving global optimization from warehouse layout to picking path selection. The key advantage of the proposed method is its ability to reduce the perceived complexity of paths. In reality, pickers typically prefer simpler routes rather than those that, while potentially optimal, contain complex or inconsistent paths. The more complex and characteristic circulation paths required pickers to exert additional effort in orientation, which not only slowed down picking speed but also increased the probability of errors. The attention mechanism in this embodiment can easily integrate such constraints by shielding related actions when constructing circulation paths. This method not only helps to simplify the picker's workflow and improve work efficiency but also effectively reduces the possibility of errors, thereby optimizing the entire picking process. This fully demonstrates that, compared with traditional methods for solving the Traveling Salesman Problem, the method proposed in this embodiment has greater flexibility and practicality in practical applications, especially in dealing with specific domain constraints.

[0189] This invention provides a warehouse operation method based on integrated optimization of storage location and picking path. The method initializes the warehouse layout based on warehouse physical structure data and pallet attribute data, generating an initial storage layout parameter set. Based on the initial storage layout parameter set, a dynamic storage location allocation optimization model is constructed with the objective of minimizing the total wasted space in the entire warehouse. The dynamic storage location allocation optimization model is iteratively solved to obtain an initial pallet storage location allocation scheme under the current warehouse layout. Responding to real-time order information, based on the initial pallet storage location allocation scheme, an attention-based picking path optimization algorithm is used to generate the optimal picking path for each order. Based on the remaining material storage status after the optimal picking path is executed, data on the remaining material layout to be optimized is obtained. The remaining material storage locations are then re-allocated based on this data to obtain a storage location reallocation scheme. Finally, an integrated storage location-path operation instruction is generated based on the optimal picking path and the storage location reallocation scheme, and the operation of the warehouse management system is controlled through this integrated storage location-path operation instruction. Compared with existing technologies, this method achieves integrated and coordinated optimization of warehouse layout, storage location allocation and picking path through the coordinated optimization of storage location allocation and picking path. This effectively reduces the total space waste of the entire warehouse, improves warehouse operation efficiency and resource utilization, and thus realizes intelligent and integrated management of warehousing operations.

[0190] It should be noted that the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0191] In one embodiment, such as Figure 11 As shown, this embodiment of the invention provides a warehouse operation system based on integrated optimization of storage location and picking path, the system comprising:

[0192] The warehouse initialization module 101 is used to initialize the warehouse layout based on warehouse physical structure data and pallet attribute data, and generate an initial warehouse layout parameter set.

[0193] The storage location model construction module 102 is used to construct a dynamic storage location allocation optimization model based on the initial storage layout parameter set, with the optimization objective of minimizing the total space waste value of the entire warehouse.

[0194] The storage location model solving module 103 is used to iteratively solve the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout.

[0195] The picking path analysis module 104 is used to respond to real-time order information and generate the optimal picking path for each order by using an attention learning-based picking path optimization algorithm based on the real-time order information and the initial pallet storage location allocation scheme.

[0196] The warehouse location reallocation module 105 is used to evaluate the remaining material layout data to be optimized based on the warehouse remaining material storage status after the optimal picking path is executed, and to perform secondary warehouse location reallocation based on the remaining material storage location to be optimized to obtain a warehouse location reallocation scheme.

[0197] The warehouse operation module 106 is used to generate integrated storage location-path operation instructions based on the optimal picking route and the storage location reallocation scheme, and control the operation of the warehouse management system through the integrated storage location-path operation instructions.

[0198] For specific limitations regarding a warehouse operation system based on integrated optimization of storage location and picking routes, please refer to the above-described limitations regarding a warehouse operation method based on integrated optimization of storage location and picking routes, which will not be repeated here. Those skilled in the art will recognize that the various modules and steps described in conjunction with the embodiments disclosed in this application can be implemented in hardware, software, or a combination of both. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0199] This invention provides a warehouse operation system based on integrated optimization of storage location and picking path. The system's storage initialization module initializes the warehouse layout based on warehouse physical structure data and pallet attribute data, generating an initial storage layout parameter set. A storage location model construction module, based on the initial storage layout parameter set, constructs a dynamic storage location allocation optimization model with the objective of minimizing total warehouse space waste. A storage location model solving module iteratively solves the dynamic storage location allocation optimization model to obtain an initial pallet storage location allocation scheme under the current warehouse layout. A picking path analysis module, responding to real-time order information, generates the optimal picking path for each order based on the initial pallet storage location allocation scheme using an attention-based picking path optimization algorithm. A storage location reallocation module obtains the remaining material layout data to be optimized based on the warehouse's remaining material storage status after the optimal picking path is executed, and performs secondary storage location allocation on the remaining material storage locations based on this data, obtaining a storage location reallocation scheme. A warehouse operation module generates integrated storage location-path operation instructions based on the optimal picking path and the storage location reallocation scheme, controlling the operation of the warehouse management system through these instructions. Compared with existing technologies, this system achieves integrated and coordinated optimization of warehouse layout, storage location allocation and picking routes through the coordinated optimization of storage location allocation and picking routes. This effectively reduces the total wasted space in the warehouse, improves warehouse operation efficiency and resource utilization, and thus realizes intelligent and integrated management of warehousing operations.

[0200] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A warehouse operation method based on integrated optimization of storage location and picking path, characterized in that, Includes the following steps: Initialize the warehouse layout based on warehouse physical structure data and pallet attribute data, and generate an initial warehouse layout parameter set; Based on the initial warehouse layout parameter set, a dynamic storage location allocation optimization model is constructed with the goal of minimizing the total wasted space value of the entire warehouse. The dynamic storage location allocation optimization model is iteratively solved to obtain the initial pallet storage location allocation scheme under the current warehouse layout; In response to real-time order information, based on the real-time order information and the initial pallet storage location allocation scheme, an attention-based picking path optimization algorithm is used to generate the optimal picking path for each order. Based on the warehouse remaining material storage status after the optimal picking route is executed, the remaining material layout data to be optimized is obtained. Based on the remaining material layout data to be optimized, the storage location of the remaining materials is re-allocated to obtain the storage location redistribution scheme. Based on the optimal picking route and the storage location reallocation scheme, an integrated storage location-path operation instruction is generated, and the operation of the warehouse management system is controlled through the integrated storage location-path operation instruction. The step of iteratively solving the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout includes: The system scans layer by layer from the bottom to the top of the warehouse. During the scanning process, it determines the set of pallets to be stored based on the initial warehouse layout parameter set and sets the current level as the bottom layer of the warehouse. The current level is taken as the level to be processed. The pallets of the level to be processed are selected from the set of pallets to be stored and sorted in descending order of pallet height to build a list of pallets to be processed. Filter out all vacant storage locations in the layer to be processed and sort them in descending order of storage location height to generate a list of available locations in the layer to be processed; Based on the list of pallets to be processed and the list of available locations, the dynamic storage location allocation optimization model is solved using a pallet location allocation algorithm to obtain a primary storage location allocation solution. Based on a local search strategy, the pallet location reallocation solution is optimized by reallocating pallet locations to obtain the initial pallet location allocation scheme under the current warehouse layout.

2. The warehouse operation method based on integrated optimization of storage location and picking path as described in claim 1, characterized in that, The step of constructing a dynamic storage location allocation optimization model based on the initialized storage layout parameter set, with the objective of minimizing the total wasted space in the entire warehouse, includes: Define a pallet stacking identifier variable and a single pallet storage location allocation variable; wherein, the single pallet storage location allocation variable is used to indicate that a single pallet is stored alone in a storage location, and the pallet stacking identifier variable is used to indicate that two different pallets are stored in the same storage location; Based on the initial warehouse layout parameter set and the predetermined waste value measurement rules, the total warehouse space waste value of all pallet storage location combinations is calculated using the pallet stacking identification variable and the single pallet storage location allocation variable. A dynamic storage location allocation optimization model is constructed with the goal of minimizing the total wasted space in the entire warehouse.

3. The warehouse operation method based on integrated optimization of storage location and picking path as described in claim 2, characterized in that, The specific process for calculating the total wasted space value of the entire warehouse is as follows: When two pallets to be stored are stored together in one storage location, the space waste generated when each pallet to be stored is stacked in the same storage location is calculated based on the sum of the heights of each pallet to be stored in the same storage location, and the double pallet stacking waste value of each pallet to be stored is obtained. When a single pallet to be stored is placed in a single pallet storage location, if it is detected that another pallet to be stored is not allowed to be placed on top of the single pallet to be stored, the restricted space waste value is obtained based on the sum of the differences between the height of all single pallet storage locations and the height of the single pallet to be stored. Based on the pre-obtained single-pallet space waste factor, measure the single-pallet space waste value when a pallet to be stored is already stored in a location that allows two pallets to be placed, and the remaining height is insufficient to place another pallet to be stored. The total space waste value of the entire warehouse is obtained by calculating the weighted sum of the waste value of double pallet stacking, the waste value of restricted space, and the waste value of single pallet space using the pallet stacking identification variable and the single pallet storage location allocation variable.

4. A warehouse operation method based on integrated optimization of storage location and picking path as described in claim 2, characterized in that: The constraints of the dynamic storage location allocation optimization model include storage restrictions for each storage location, pallet storage feasibility constraints, space restrictions, hierarchical restrictions, and stacking rule restrictions. The storage restriction for each storage location is specifically that the number of pallets stored in each storage location shall not exceed its maximum capacity limit; The pallet storage feasibility constraint specifically states that each pallet can only be stored in one storage location; The space constraint specifically states that the total height of all pallets stored in each storage location shall not exceed the preset maximum storage location height. The hierarchical restriction constraint specifically means that the number of pallet storage layers shall not exceed its maximum allowed number of layers; The stacking rule constraint specifically states that if neither of the two pallets is allowed to place another pallet on top of it, then the two pallets cannot be stored in the same storage location.

5. A warehouse operation method based on integrated optimization of storage location and picking path as described in claim 1, characterized in that, The step of solving the dynamic storage location allocation optimization model using a pallet location allocation algorithm based on the list of pallets to be processed and the list of available locations to obtain the initial storage location allocation solution includes: The set of unstored pallets is initialized to an empty set, and when the list of pallets to be processed is not empty, the pallet allocation operation is performed according to the constraints of the dynamic storage location allocation optimization model, and the pallets in the list of pallets to be processed are checked in turn. If a pallet with a height greater than the height of the first storage location in the available location list is detected in the pallet list to be processed, it is moved from the pallet list to the set of unstored pallets, and the detection of the next pallet in the pallet list to be processed continues. Select the tallest pallet from the list of pallets to be processed as the reference pallet. Then, search for the second tallest pallet in the list of pallets to be processed. If the sum of the heights of the reference pallet and the second tallest pallet does not exceed the height of the first storage location in the list of available locations, and at least one pallet allows another pallet to be placed on top of it, then store the reference pallet and the second tallest pallet as paired pallets in the lowest available storage location; otherwise, only store the reference pallet in the lowest available storage location. If the selected lowest available storage location is a double-depth storage location, then search the list of pallets to be processed for other stackable pallets or pallet pairs to fill the remaining space of the double-depth storage location. Remove the selected paired pallets from the list of pallets to be processed and the set of pallets to be stored, and remove the used available storage locations from the list of available locations; If the unstored pallet set is empty and the processing level is not the lowest level, then the pallets in the unstored pallet set are re-inserted into the processing pallet list in descending order of pallet height, the current level is adjusted to the previous level, and the available position list is regenerated. If the set of unstored pallets is empty and the level to be processed is the lowest level, select the pallet with the smallest height from the set of unstored pallets and store it in the remaining empty storage space of the current level or the level below. If the current level is less than the maximum level and the set of unstored pallets is empty, then update the level to be processed to the next level, and re-filter the list of pallets to be processed and available locations for the current level, and continue to allocate pallets until all pallets meet the constraints of the dynamic storage location allocation optimization model, then stop the iteration and output the initial storage location allocation solution.

6. A warehouse operation method based on integrated optimization of storage location and picking path as described in claim 1, characterized in that, The step of optimizing the pallet location reallocation scheme of the primary storage location allocation solution based on a local search strategy to obtain the initial pallet storage location allocation scheme under the current warehouse layout includes: Based on the storage location type and pallet stacking status, calculate the storage location waste value of each storage location in the primary storage location allocation solution, and sort all storage locations in the primary storage location allocation solution in descending order according to the storage location waste value to generate a list of storage locations to be optimized; Traverse the list of storage locations to be optimized, and take each storage location in the list of storage locations to be optimized as the source location, and detect feasible moves in sequence according to the four pre-built priority of move types; If a feasible move is detected, the feasible move is executed in descending order of waste reduction magnitude, and the pallet position and storage position status are updated in real time based on the storage location information of the feasible move. After each feasible move is executed, the total waste value of the entire warehouse space is recalculated. If the total waste value of the entire warehouse space is greater than the preset waste threshold, a new list of storage locations to be optimized is generated based on the updated storage location status. The above steps are repeated to continuously iterate and optimize until the total waste value of the entire warehouse space is less than the preset waste threshold. Then, the iterative optimization is terminated, and the initial pallet storage location allocation scheme under the current warehouse layout is output.

7. A warehouse operation method based on integrated optimization of storage location and picking path as described in claim 1, characterized in that, The step of generating the optimal picking route for each order using an attention-based picking route optimization algorithm based on the real-time order information and the initial pallet storage location allocation scheme includes: The list of pallets to be picked is obtained based on the real-time order information, and the current warehouse layout and pallet storage location information are extracted from the initial pallet storage location allocation scheme based on the list of pallets to be picked. Based on the pallet storage location information, the current warehouse layout is converted into a graph structure to obtain a warehouse topology graph; wherein, the graph nodes of the warehouse topology graph include pallet storage location nodes, shelf aisle top and bottom nodes, and warehouse entrance and exit nodes, and the edges represent the traversable path relationships between graph nodes; An attention mechanism is used to learn the cyclic vertical edges and cyclic horizontal edges from the warehouse topology graph, and a cyclic graph is constructed according to the order of the cyclic vertical edges and the cyclic horizontal edges. The attention mechanism includes an input layer, an embedding layer, a multi-head attention mechanism layer, a feedforward layer, a residual connection and layer normalization layer, and a linear layer. Extract all valid picking routes from the warehouse entrance node, through each pallet storage location node, and to the warehouse exit node from the aforementioned circulation map, and select the valid picking route with the minimum total cost as the optimal picking path.

8. A warehouse operation method based on integrated optimization of storage location and picking path as described in claim 1, characterized in that, The step of performing secondary allocation of storage locations for remaining materials based on the remaining material layout data to be optimized, and obtaining a storage location reallocation scheme, includes: Based on the remaining material layout data to be optimized, a storage and movement list is generated, and the maximum allowed execution time is defined; The storage move operations in the storage move list are executed sequentially according to the preset move priority, and a random number is generated during the execution of each storage move. The number of storage moves equal to the number of random numbers is randomly selected from the storage move list for deletion. During each storage move, the execution time of each storage move is calculated and accumulated. When the accumulated execution time reaches the maximum allowed execution time, the storage move is stopped, and the currently executed storage move operations are combined to form a storage move operation scheme. Based on the total waste reduction of the storage relocation operation scheme, the optimal storage location reallocation scheme is selected from all storage relocation operation schemes.

9. A warehouse operation system based on integrated optimization of storage location and picking path, characterized in that, The warehouse operation method based on integrated optimization of storage location and picking path as described in claim 1, wherein the system comprises: The warehouse initialization module is used to initialize the warehouse layout based on warehouse physical structure data and pallet attribute data, and generate an initial warehouse layout parameter set. The storage location model construction module is used to construct a dynamic storage location allocation optimization model based on the initial storage layout parameter set, with the optimization objective of minimizing the total space waste value of the entire warehouse. The storage location model solving module is used to iteratively solve the dynamic storage location allocation optimization model to obtain the initial pallet storage location allocation scheme under the current warehouse layout. The picking route analysis module is used to respond to real-time order information and generate the optimal picking route for each order based on the real-time order information and the initial pallet storage location allocation scheme using an attention learning-based picking route optimization algorithm. The warehouse location reallocation module is used to evaluate the remaining material layout data to be optimized based on the warehouse remaining material storage status after the optimal picking path is executed, and to perform secondary warehouse location reallocation based on the remaining material storage location to be optimized to obtain a warehouse location reallocation scheme. The warehouse operation module is used to generate integrated storage location-path operation instructions based on the optimal picking route and the storage location reallocation scheme, and control the operation of the warehouse management system through the integrated storage location-path operation instructions.

Images