Order-driven mixed box stacking method, system and equipment and medium

Abstract

The invention discloses an order-driven mixed box stacking method, system and equipment and a medium. According to the method, a stacking process is decoupled into two associated subtasks of active warehouse-out decision and real-time stacking execution; the problems of frequent stack transfer and invalid reciprocating carrying caused by scattered goods combination in a physical layout limited scene in an existing offline stacking algorithm are effectively solved, and meanwhile, the space utilization rate bottleneck caused by lack of order global view in a traditional online stacking algorithm is also overcome; and a brand-new framework support and a global optimization decision view angle are provided for efficient disassembly operation in a limited physical space. An energy-based warehouse-out sequence optimization algorithm is introduced into a warehouse-out decision-making layer, and the algorithm deeply fuses the local fine sampling capability of Langevin dynamics and the global evolution capability of a genetic algorithm, so that the technical defects that a traditional heuristic rule is difficult to capture complex geometric features of goods and is easy to fall into local optimum are overcome; and the three-dimensional warehouse is guided to realize active and efficient replenishment scheduling.

Description

Technical Field

[0001] This application belongs to the field of warehouse palletizing technology, specifically relating to an order-driven mixed-box palletizing method, system, equipment, and medium. Background Technology

[0002] Against the backdrop of the deepening development of the new retail warehousing model, and with the explosive growth of personalized consumer demand, logistics systems face the severe challenge of processing fragmented and split orders at high frequency. As a core component of the warehouse scheduling system, the efficiency of solving the Three-Dimensional Bin Packing Problem (3D-BPP) directly determines the utilization rate of warehouse space, order processing speed, and overall operating costs. Currently, solutions to this problem mainly rely on traditional heuristic algorithms, offline packing algorithms based on deep reinforcement learning, and online packing algorithms.

[0003] However, existing technologies have significant limitations in complex new retail warehousing scenarios. In such scenarios, such as... Figure 1 As shown, the core operation is a real-time depalletizing and palletizing process. The robotic arm must grab boxes from the pallets to be depalletized from the automated warehouse and immediately place them onto the pallets to be palletized. This process strictly adheres to the principle of immediate palletizing of each depalletized box, and temporary storage, retraction, or secondary depalletizing is strictly prohibited. Due to the physical layout of the work site, the work area typically only holds one pallet to be depalletized and one pallet to be palletized. This stringent capacity constraint makes it difficult for existing algorithms to effectively adapt to this scenario. Specifically, while heuristic and offline palletizing algorithms can optimize order sequences from a global perspective, their generated loading schemes often contain combinations of goods scattered across different pallets to be depalletized. With the work area limited to a single pallet, this easily leads to frequent rearrangement of pallets to be depalletized and ineffective back-and-forth handling, significantly increasing operational complexity and time costs. On the other hand, online palletizing algorithms rely solely on locally arriving box information for decision-making, lacking proactive selection of the goods' outbound order. This lack of a global perspective results in a bottleneck in space utilization. Therefore, the traditional 3D-BPP framework is difficult to implement effectively in complex and dynamic new retail scenarios, and there is an urgent need to build a mixed-box palletizing framework that takes into account both global order optimization and real-time box response. Summary of the Invention

[0004] To address the warehousing and logistics bottlenecks caused by the highly fragmented and individually distributed orders under the new retail model, this application proposes an order-driven mixed-box palletizing method, system, equipment, and medium. By decoupling the traditional palletizing process into two related sub-tasks—proactive outbound decision-making and real-time palletizing execution—it provides a new architectural support and a globally optimized decision-making perspective for efficient individually distributed palletizing operations under limited physical space.

[0005] This application is achieved through the following technical solution:

[0006] An order-driven mixed-box palletizing method includes:

[0007] Initialization: Sort the order set according to the weight of the goods to generate an initial order queue, and initialize the sets of pallets to be unpacked, pallets to be stacked, and completed pallets to an empty set; the order queue includes several types of goods identifiers and their attribute vectors;

[0008] Outbound replenishment: When the pallet to be unpacked is in an idle state, the geometric space state of the current pallet to be unpacked is analyzed by an energy-based outbound order optimization algorithm, the optimal matching goods are selected from the order queue, an outbound instruction is sent to guide the automated warehouse to transfer the box corresponding to the optimal matching goods to the pallet to be unpacked, and the outbound goods are removed from the order queue.

[0009] Decomposition and palletizing: Real-time sensing of the geometric dimensions of the boxes on the pallet to be decomposition, and using an online palletizing algorithm to plan the optimal placement position and orientation of the boxes on the current pallet to be palletized based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be palletized, and sending corresponding commands to the handling unit;

[0010] Return to warehouse reset: Once the best-matched goods meet the order splitting requirements, the remaining boxes on the pallet to be split will be transferred back to the automated warehouse, and the pallet to be split will be restored to an idle state to automatically trigger the next round of outbound replenishment process. Outbound replenishment, splitting and palletizing, and return to warehouse reset are executed in a loop until the order queue is completely cleared and the order task ends.

[0011] In some implementations, the inputs to the outbound order optimization algorithm include the order queue, the number of iterations, the population size, the temperature parameter, and the step size factor, and the output is the optimal outbound sequence selected from the solution space;

[0012] The outbound sequence optimization algorithm constructs an energy function that quantifies space utilization and models the decision optimization process of candidate outbound sequences as an energy minimization problem.

[0013] A hybrid optimization strategy combining local sampling with a Langevin update operator and global evolution with a genetic algorithm is adopted to retrieve the optimal outbound sequence from the candidate outbound sequences.

[0014] In some implementations, the energy function is expressed as:

[0015] ;

[0016] in, Indicates candidate outbound sequence Energy; This indicates the space utilization rate.

[0017] In some implementations, two versions of space utilization assessment are used:

[0018] Online version, This indicates the space utilization rate after simulating palletizing of the current pallet to be stacked;

[0019] Offline version This represents the overall space utilization rate after performing a complete simulated palletizing of all goods in the order queue.

[0020] In some implementations, the hybrid optimization strategy of using an integrated Langevin update operator for local sampling and a genetic algorithm for global evolution to retrieve the optimal outbound sequence from candidate outbound sequences includes:

[0021] A population is initialized based on the population size. The first individual in the population is the order queue, and the remaining individuals are randomly generated from the order queue. Each individual represents a candidate outbound sequence.

[0022] Local optimization: For any candidate outbound sequence, the Langevin update operator generates a neighborhood solution by randomly swapping the outbound order of two goods in the sequence, and calculates the energy difference between the energy corresponding to the neighborhood solution and the energy corresponding to the original candidate outbound sequence; based on the energy difference, temperature parameters, and step size factor, the probability of an individual accepting a new solution is determined, and the individual state is updated according to this probability;

[0023] Global evolution: After all candidate outgoing sequences have completed the local optimization of the Langevin update operator, the genetic algorithm performs the survival of the fittest selection logic based on the energy value of the individuals, selects the superior parent individuals from the current population for subsequent evolution, and the selected parent individuals generate offspring through the sequential crossover operator, and the offspring genes are fine-tuned by the mutation operator.

[0024] The optimal outbound sequence is retrieved by iteratively optimizing the local and global processes described above.

[0025] In some implementations, the method of using an online palletizing algorithm to plan the optimal placement position and orientation of the box on the current pallet based on the box's geometric dimensions and the three-dimensional state information of the current pallet to be palletized further includes:

[0026] If the current waiting pallet is insufficient to accommodate the boxes, the current waiting pallet is stored in the completed pallet set and a new empty waiting pallet is created to continue the operation until the splitting requirement of the order is met.

[0027] Secondly, this application proposes an order-driven mixed-box palletizing system for implementing the mixed-box palletizing method described in any of the above embodiments, comprising:

[0028] The input module is configured to: input order tasks and generate corresponding order sets;

[0029] The initialization module is configured to sort the order set according to the weight of the goods, generate an initial order queue, and initialize the set of pallets to be unpacked, pallets to be stacked, and completed pallets to an empty set.

[0030] The dynamic outbound module is configured to: when the pallet to be unloaded is idle, use an energy-based outbound order optimization algorithm to select the best matching goods from the order queue, send an outbound instruction to guide the automated warehouse to transfer the box corresponding to the best matching goods to the pallet to be unloaded, and remove the outbound goods from the order queue;

[0031] In addition, the online palletizing module is configured to: perceive the geometric dimensions of the boxes on the pallet to be unloaded in real time, and based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be stacked, plan the optimal placement position and orientation of the boxes on the current pallet to be stacked and send corresponding commands to the operation execution module.

[0032] In some implementations, the operation execution module includes: an automated warehouse, pallets to be unpacked, pallets to be stacked, and a handling unit connecting the various areas;

[0033] The automated warehouse releases the corresponding goods and transfers them to the pallets to be unloaded according to the instructions sent by the dynamic outbound module.

[0034] The handling unit, in the area to be unpacked, executes the command issued by the online palletizing module to grab the box and place it on the pallet to be stacked;

[0035] Once the target boxes on the pallet to be unloaded have been removed or the order requirements have been met, the remaining boxes are transferred back to the warehouse by the handling unit, triggering a closed-loop feedback mechanism to enter the next round of decision-making.

[0036] Thirdly, this application proposes an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the above-described mixed-box palletizing methods.

[0037] Fourthly, this application proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described embodiments of the mixed-box palletizing method.

[0038] This application proposes an order-driven mixed-box palletizing method, decoupling the palletizing process into two related sub-tasks: proactive outbound decision-making and real-time palletizing execution. This effectively solves the problem of frequent repackaging and ineffective reciprocating handling caused by the dispersed cargo combinations in existing offline palletizing algorithms under physically constrained scenarios. It also overcomes the space utilization bottleneck caused by the lack of a global order perspective in traditional online palletizing algorithms, providing a new architectural support and a global optimization decision-making perspective for efficient splitting operations under limited physical space. Furthermore, an energy-based outbound sequence optimization algorithm is introduced at the outbound decision layer: by constructing an energy function that quantifies space utilization and modeling the sequence decision as an energy minimization problem, a scientific and intuitive quantitative evaluation of the outbound sequence is achieved. This algorithm deeply integrates the local fine-sampling capability of Langevin dynamics with the global evolution capability of genetic algorithms, overcoming the technical shortcomings of traditional heuristic rules that struggle to capture complex geometric features of goods and are prone to getting trapped in local optima. It can quickly and accurately select the cargo sequence that best matches the remaining space of the current pallet from a large number of candidate orders, thereby guiding the automated warehouse to achieve proactive and efficient replenishment scheduling.

[0039] Accordingly, the order-driven mixed-box palletizing system, electronic device, and computer-readable storage medium proposed in this application also possess the same technical effects as described above. Attached Figure Description

[0040] The accompanying drawings, which are included to provide a further understanding of the embodiments of this application and form part of this application, do not constitute a limitation on the embodiments of this application. In the drawings:

[0041] Figure 1 This is a schematic diagram of a new retail warehousing environment.

[0042] Figure 2 This is a flowchart of the mixed-box palletizing method proposed in the embodiments of this application;

[0043] Figure 3 A visual comparison of palletizing strategies for different outbound decision-making strategies under the same order in this application embodiment;

[0044] Figure 4 This is a block diagram illustrating the principle of the mixed-box palletizing system proposed in an embodiment of this application.

[0045] Figure 5 This is a schematic diagram of the electronic device proposed in the embodiments of this application;

[0046] Figure 6 This is a schematic diagram of a computer-readable storage medium proposed in an embodiment of this application.

[0047] Figure reference numerals and corresponding component names:

[0048] 200-Mixed box palletizing system, 201-Input module, 202-Initialization module, 203-Dynamic outbound module, 204-Online palletizing module, 205-Job execution module, 300-Electronic device, 310-Memory, 320-Processor, 311-Computer program A, 400-Computer-readable storage medium, 411-Computer program B. Detailed Implementation

[0049] In the following, the terms “comprising” or “may include” as used in the various embodiments of this application indicate the presence of a function, operation, or element of the invention and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in the various embodiments of this application, the terms “comprising,” “having,” and their cognates are intended only to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or adding one or more combinations of the foregoing.

[0050] In various embodiments of this application, the expression "or" or "at least one of A and / or B" includes any combination or all combinations of the words listed simultaneously. For example, the expression "A or B" or "at least one of A and / or B" may include A, may include B, or may include both A and B.

[0051] The terms used in the various embodiments of this application (such as "first," "second," etc.) may modify various constituent elements in the various embodiments, but do not limit the corresponding constituent elements. For example, the above terms do not limit the order and / or importance of the elements. The above terms are only used for the purpose of distinguishing one element from other elements. For example, a first user device and a second user device refer to different user devices, although both are user devices. For example, without departing from the scope of the various embodiments of this application, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element.

[0052] It should be noted that if a description is made of "connecting" one component to another, then the first component can be directly connected to the second component, and a third component can be "connected" between the first and second components. Conversely, when a component is "directly connected" to another component, it can be understood that there is no third component between the first and second components.

[0053] The terminology used in the various embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the various embodiments of this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. The terms (such as those defined in a generally used dictionary) are to be interpreted as having the same meaning as in the context of the relevant technical field and are not to be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the embodiments and accompanying drawings. The illustrative embodiments and descriptions of this application are only for explaining this application and are not intended to limit this application.

[0055] This application proposes an order-driven split-palletizing (ODSP) method, which decouples the traditional palletizing process into two related sub-tasks: proactive outbound decision-making and real-time palletizing execution. This effectively addresses the bottlenecks in operational efficiency caused by the highly fragmented nature of orders in the new retail model.

[0056] Specifically, such as Figure 2 As shown, the mixed-box palletizing method proposed in this application includes the following steps:

[0057] Step 1, Initialization: Sort the order set according to the weight of the goods, generate an initial order queue, and initialize the set of pallets to be unpacked, pallets to be stacked, and completed pallets to an empty set;

[0058] Step 2, Outbound Replenishment: When the pallet to be unpacked is idle, the energy-based outbound order optimization algorithm is used to analyze the geometric space state of the current pallet to be unpacked, select the best matching goods from the order queue, send outbound instructions to guide the automated warehouse to transfer the box corresponding to the best matching goods to the pallet to be unpacked, and remove the outbound goods from the order queue.

[0059] Step 3, Decomposition and Palletizing: Real-time sensing of the geometric dimensions of the boxes on the pallet to be decomposition, and using an online palletizing algorithm to plan the optimal placement position and orientation of the boxes on the current pallet to be palletized based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be palletized, and sending corresponding commands to the handling unit.

[0060] Step 4, Reset to Warehouse: Once the best-matched goods meet the order splitting requirements, the remaining boxes on the pallet to be split will be transferred back to the automated warehouse, and the pallet to be split will be restored to an idle state so as to automatically trigger the next round of outbound replenishment process. Steps 2-4 are executed in a loop until the order queue is completely cleared and the order task ends.

[0061] Furthermore, in step 1 of this application embodiment, the order set is defined as... ;in; Indicates the first The attribute vector of the class of goods; where, This indicates the quantity of the goods required for individual shipments. , These represent the length, width, height, and weight of the goods, respectively. Indicates the type of goods. Based on the weight of the goods. Sort the order set in descending order to generate the initial order queue. ;in, Arrange the mapping index. Simultaneously, the tray to be unpacked... pallets awaiting coding and completed pallet collection All are initialized to an empty set.

[0062] Furthermore, in step 2 of this embodiment, the Energy-Based Outbound Sequence Optimization (EBSO) algorithm guides the optimal outbound logic of the automated warehouse by constructing an energy function that quantifies space utilization. The input to this algorithm includes the order queue. Number of iterations Population size Temperature parameters and step size factor The output is the optimal outbound sequence selected from the solution space. The EBSO algorithm selects candidate outbound sequences. The decision-making process is modeled as an energy minimization problem, and a population is initialized. ,in The remaining individuals are generated by random perturbations; Represents the order queue number There are 10 candidate outbound sequences. Since the outbound sequences have discrete symbol arrangement characteristics and the search space does not have continuous gradients, the algorithm adopts a hybrid optimization strategy that integrates the Langevin Update Operator (LUO) for local sampling and the Genetic Algorithm (GA) for global evolution, thereby achieving efficient retrieval of the optimal outbound sequence.

[0063] The energy function is defined as follows: ,in, It uses a built-in deep reinforcement learning policy network The predicted space utilization rate is obtained after simulated palletizing. This energy function establishes the physical correlation between outbound decisions and the underlying palletizing operations; the lower the energy value, the higher the geometric feature matching degree between the sequence and the current pallet to be palletized. For different industrial application needs, this application's embodiments employ two evaluation methods. Version: In the online version (EBSO-online), This represents the space utilization rate after simulated palletizing of the current pallet to be stacked. Whenever the system triggers a pallet replacement instruction, the EBSO algorithm is immediately invoked to accurately capture the geometric fit between the pallet to be stacked and the goods to be shipped, thereby guiding the automated warehouse in shipping. The offline version (EBSO-offline) will... Defined as the overall space utilization rate after performing a complete simulated palletizing of all goods in the order queue. It is important to emphasize that in offline mode, the optimal outbound sequence is determined through the EBSO algorithm during the task initialization phase and strictly executed in subsequent workflows, without dynamic reordering.

[0064] Furthermore, the built-in deep reinforcement learning policy network This utilizes existing mainstream models such as PCT and GOPT. It employs these mainstream 3D-BPP deep reinforcement learning models to predict and output the optimal space utilization of boxes within the pallet to be unpacked on the pallet to be stacked. The specific technical principle is manifested in: a policy network. The input typically includes a 3D representation of the current palletizing environment, i.e., the 3D state information of the pallet to be palletized (e.g., the remaining available space of the pallet represented by a height map, multi-dimensional point cloud, or 3D raster), and geometric attribute vectors of candidate boxes to be placed (e.g., physical attribute parameters such as length, width, height, and weight). The network output is the placement action instruction, i.e., the policy network predicts and outputs the optimal placement position coordinates of the target box in the 3D coordinate system of the current pallet to be palletized based on the current input state. and spatial rotation direction The aim is to maximize the physical stability and space compactness of stacking operations.

[0065] The LUO operator is responsible for implementing local fine-grained sampling of individuals in the population, exploring the neighborhood space of the current outgoing sequence by introducing a random perturbation mechanism. For any candidate outgoing sequence... The LUO operator generates neighborhood solutions by randomly swapping the shipping order of two goods in a sequence. And calculate the energy difference between the two. To effectively avoid local extrema traps common in combinatorial optimization problems, the LUO operator strictly follows the Metropolis-Hastings probability criterion to update individual states, and its probability of accepting a new solution is defined as follows: This mechanism gives the algorithm the ability to accept inferior solutions with higher energy levels with a certain probability in the early stages of optimization, simulating the thermal motion process in physical annealing, thereby ensuring that the sequence search can robustly converge to low-energy states.

[0066] The GA algorithm works in conjunction with the LUO operator to further improve search depth and population diversity through swarm intelligence. After each round of local optimization using LUO, the GA algorithm calculates the energy value of each individual element. The selection process employs a survival-of-the-fittest logic, choosing superior parent individuals from the current population for subsequent evolution. The selected parent individuals generate offspring through sequential crossover, and the offspring's genes are fine-tuned using mutation operators, thereby maintaining population diversity while continuously identifying and optimizing the release strategy.

[0067] Furthermore, in step 3 of this embodiment, the online palletizing algorithm is primarily responsible for handling real-time placement tasks under the stringent constraint of strong coupling between piecework and palletizing. It employs a deep reinforcement learning policy network. (This embodiment can employ 3D-BPP deep learning architectures such as PCT, GOPT, or SOP) to achieve real-time placement decisions under the strong coupling constraints of piecework and palletizing. Specifically, this policy network... The real-time sensed geometric dimension vector of the container to be chained and the current three-dimensional spatial state features of the pallet to be chained are used as joint inputs. The feature encoding operator within the network extracts the spatial correlation features between the geometric features of the goods and the remaining available space on the pallet. Subsequently, the policy network... Based on the pre-trained optimal policy mapping, probabilistic inference is performed in the candidate action space to predict and output the optimal projection coordinates of the target box in the coordinate system of the pallet to be encoded. and spatial rotation angle It enables precise planning of the placement and orientation of containers within milliseconds, and maximizes the space compactness and loading efficiency of pallets while ensuring the physical stability of stacking.

[0068] If the tray is to be labeled If the current pallet is insufficient to accommodate the box, then the pallet to be stacked will be moved. Store completed tray collection Then create new empty pallets to continue operations until the splitting requirements for this batch of goods are met.

[0069] The mixed-box palletizing method proposed in this application decouples the traditional palletizing process into two related sub-tasks: proactive outbound decision-making and real-time palletizing execution. This effectively solves the problem of frequent repackaging and ineffective reciprocating handling caused by the dispersed combination of goods in physically constrained scenarios, which is a challenge of existing offline palletizing algorithms. It also overcomes the space utilization bottleneck caused by the lack of a global order perspective in traditional online palletizing algorithms, providing a new architectural support and a global optimization decision-making perspective for efficient piecework operations in constrained physical spaces. At the core logic level of dynamic outbound decision-making, an energy-based outbound sequence optimization algorithm is innovatively introduced: by constructing an energy function that quantifies space utilization and modeling the sequence decision as an energy minimization problem, a scientific and intuitive quantitative evaluation of the outbound sequence is achieved. This algorithm deeply integrates the local fine-sampling capability of Langevin dynamics and the global evolution capability of genetic algorithms, solving the technical shortcomings of traditional heuristic rules that are difficult to capture the complex geometric features of goods and are prone to getting trapped in local optima. It can quickly and accurately select the goods sequence that best matches the remaining space of the current pallet from a large number of candidate orders, thereby guiding the automated warehouse to achieve proactive and efficient replenishment scheduling.

[0070] Integrating the EBSO algorithm into mainstream online palletizing algorithms, its space utilization is significantly better than that of traditional manual heuristic rules, as shown in Table 1 and... Figure 3 As shown, taking the SOP algorithm as an example, the average space utilization (Uti) after adopting the EBSO-online strategy reaches 0.725, an improvement of approximately 8.8% compared to manual rules. Comparing the two versions, EBSO-online and EBSO-offline, the former is slightly better than the latter in terms of average space utilization. For example, under the GOPT algorithm, the Uti of EBSO-online reaches 0.711, while that of EBSO-offline is 0.709. This indicates that EBSO-online can more accurately select the matching order of goods shipments by capturing the dynamic geometric features of the remaining space of the pallet to be stacked in real time, thereby achieving a more compact palletizing distribution. This result strongly demonstrates that in complex split-palletizing scenarios, relying solely on traditional manual heuristic rules is insufficient to cope with the diversity of goods geometric characteristics, but the palletizing potential can be improved through order optimization.

[0071] The average space utilization rate is defined as: (previous) Individual code fills the tray space effect The arithmetic mean, i.e. .

[0072] Table 1. Performance metrics comparison of integrating different outbound strategies with pre-pile palletizing algorithms

[0073]

[0074] As shown in Table 1, this application improves decision-making optimization capabilities by introducing the EBSO algorithm without substantially increasing the algorithm's execution time, achieving an effective balance between search accuracy and computational efficiency. Furthermore, EBSO-online exhibits superior dynamic robustness. In actual operations, if abnormal interference such as palletizing failure or placement deviation occurs, EBSO-online can adjust the subsequent outbound order in real time based on the pallet status feedback. In contrast, EBSO-offline relies solely on static planning at the initial moment and lacks the ability to resist interference and correct errors in response to dynamic environmental changes.

[0075] Furthermore, to support the implementation of the above-mentioned mixed-box palletizing method, embodiments of this application also propose an order-driven mixed-box palletizing system, such as... Figure 4 As shown, the mixed-box palletizing system 200 includes:

[0076] Input module 201 is configured to input order tasks and generate corresponding order sets. The specific definition of the order set is as described in the mixed-box palletizing method above, and will not be repeated here.

[0077] Initialization module 202 is configured to sort the order set according to the weight of the goods, generate an initial order queue, and initialize the sets of pallets to be unpacked, pallets to be stacked, and completed pallets to an empty set. The specific initialization process is as described in the mixed palletizing method above, and will not be repeated here.

[0078] The dynamic outbound module 203 is configured to: when the pallet to be unloaded is idle, use an energy-based outbound sequence optimization algorithm to select the optimal matching goods from the order queue, send an outbound instruction to guide the automated warehouse to transfer the box corresponding to the optimal matching goods to the pallet to be unloaded, and remove the outbound goods from the order queue. The specific implementation process of the outbound sequence optimization algorithm is as described in the mixed box palletizing method above, and will not be repeated here.

[0079] Furthermore, the online palletizing module 204 is configured to: perceive the geometric dimensions of the boxes on the pallet to be unloaded in real time, and based on the box geometry and the three-dimensional state information of the current pallet to be stacked, plan the optimal placement position and orientation of the boxes on the current pallet to be stacked, and issue corresponding commands to the job execution module 205. The specific online palletizing algorithm is as described in the mixed box palletizing method above, and will not be repeated here.

[0080] Furthermore, the operation execution module 205 in this embodiment is configured to respond to and execute instructions and commands issued by the dynamic outbound module 203 and the online palletizing module 204. Specifically, the operation execution module 205 includes an automated warehouse, a pallet to be unloaded, a pallet to be stacked, and a handling unit connecting the various areas. The automated warehouse, based on the optimal outbound instruction sent by the dynamic outbound module 203, outbounds the corresponding goods and transfers them to the pallet to be unloaded. The handling unit may employ, but is not limited to, a robotic arm, which, based on the commands issued by the online palletizing module 204 in the unloaded area, grabs boxes and stacks them onto the pallet to be stacked. When all the target boxes on the pallet to be unloaded are retrieved or the order requirements are met, the remaining boxes are transferred back to the warehouse by the handling unit, and the system then triggers a closed-loop feedback mechanism to enter the next round of decision-making.

[0081] Based on the same technical concept described above, this application also proposes an electronic device, such as... Figure 5 As shown, the electronic device 300 includes: a memory 310, a processor 320, and a computer program A311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program A311, it performs the following steps:

[0082] Initialization: Sort the order set according to the weight of the goods, generate an initial order queue, and initialize the sets of pallets to be unpacked, pallets to be stacked, and completed pallets to empty sets;

[0083] Outbound replenishment: When the pallet to be unpacked is idle, the energy-based outbound order optimization algorithm is used to analyze the geometric space state of the current pallet to be unpacked, select the best matching goods from the order queue, send outbound instructions to guide the automated warehouse to transfer the box corresponding to the best matching goods to the pallet to be unpacked, and remove the outbound goods from the order queue.

[0084] Decomposition and palletizing: Real-time sensing of the geometric dimensions of the boxes on the pallet to be decomposition, and using an online palletizing algorithm to plan the optimal placement position and orientation of the boxes on the current pallet to be palletized based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be palletized, and sending corresponding commands to the handling unit;

[0085] Return to warehouse reset: Once the best-matched goods meet the order splitting requirements, the remaining boxes on the pallet to be split will be transferred back to the automated warehouse, and the pallet to be split will be restored to an idle state so as to automatically trigger the next round of outbound replenishment process. Outbound replenishment, splitting and palletizing and return to warehouse reset are executed in a loop until the order queue is completely cleared and the order task ends.

[0086] Optionally, when the processor 320 executes the computer program A311, it can implement any of the embodiments in the corresponding examples of the above-described mixed box palletizing method.

[0087] It should be noted that the electronic device proposed in this application embodiment is a device used to implement the above-mentioned mixed box palletizing method. Therefore, based on the above-mentioned mixed box palletizing method proposed in this application embodiment, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this application embodiment. Therefore, how the electronic device specifically implements the above-mentioned mixed box palletizing method will not be described in detail here. Any electronic device used by those skilled in the art to implement the above-mentioned mixed box palletizing method is within the scope of protection of this application.

[0088] Based on the same technical concept described above, embodiments of this application also propose a computer-readable storage medium, such as... Figure 6 As shown, the computer-readable storage medium 400 stores a computer program B411, which, when executed by a processor, performs the following steps:

[0089] Initialization: Sort the order set according to the weight of the goods, generate an initial order queue, and initialize the sets of pallets to be unpacked, pallets to be stacked, and completed pallets to empty sets;

[0090] Outbound replenishment: When the pallet to be unpacked is idle, the energy-based outbound order optimization algorithm is used to analyze the geometric space state of the current pallet to be unpacked, select the best matching goods from the order queue, send outbound instructions to guide the automated warehouse to transfer the box corresponding to the best matching goods to the pallet to be unpacked, and remove the outbound goods from the order queue.

[0091] Decomposition and palletizing: Real-time sensing of the geometric dimensions of the boxes on the pallet to be decomposition, and using an online palletizing algorithm to plan the optimal placement position and orientation of the boxes on the current pallet to be palletized based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be palletized, and sending corresponding commands to the handling unit;

[0092] Return to warehouse reset: Once the best-matched goods meet the order splitting requirements, the remaining boxes on the pallet to be split will be transferred back to the automated warehouse, and the pallet to be split will be restored to an idle state so as to automatically trigger the next round of outbound replenishment process. Outbound replenishment, splitting and palletizing and return to warehouse reset are executed in a loop until the order queue is completely cleared and the order task ends.

[0093] Optionally, when the computer program B411 is executed by the processor, it can implement any of the embodiments corresponding to the above-described mixed box palletizing method.

[0094] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0095] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0096] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0097] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0098] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0099] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An order-driven mixed-box palletizing method, characterized in that, include: Initialization: Sort the order set according to the weight of the goods to generate an initial order queue, and initialize the sets of pallets to be unpacked, pallets to be stacked, and completed pallets to an empty set; the order queue includes several types of goods identifiers and their attribute vectors; Outbound replenishment: When the pallet to be unpacked is in an idle state, the geometric space state of the current pallet to be unpacked is analyzed by an energy-based outbound order optimization algorithm, the optimal matching goods are selected from the order queue, an outbound instruction is sent to guide the automated warehouse to transfer the box corresponding to the optimal matching goods to the pallet to be unpacked, and the outbound goods are removed from the order queue. Decomposition and palletizing: Real-time sensing of the geometric dimensions of the boxes on the pallet to be decomposition, and using an online palletizing algorithm to plan the optimal placement position and orientation of the boxes on the current pallet to be palletized based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be palletized, and sending corresponding commands to the handling unit; Return to warehouse reset: Once the best-matched goods meet the order splitting requirements, the remaining boxes on the pallet to be split will be transferred back to the automated warehouse, and the pallet to be split will be restored to an idle state to automatically trigger the next round of outbound replenishment process. Outbound replenishment, splitting and palletizing, and return to warehouse reset are executed in a loop until the order queue is completely cleared and the order task ends.

2. The order-driven mixed-box palletizing method according to claim 1, characterized in that, The inputs to the outbound order optimization algorithm include the order queue, the number of iterations, the population size, the temperature parameter, and the step size factor. The output is the optimal outbound sequence selected from the solution space. The outbound sequence optimization algorithm models the decision optimization process of candidate outbound sequences as an energy minimization problem by constructing an energy function that quantifies space utilization. A hybrid optimization strategy combining local sampling with a Langevin update operator and global evolution with a genetic algorithm is adopted to retrieve the optimal outbound sequence from the candidate outbound sequences.

3. The order-driven mixed-box palletizing method according to claim 2, characterized in that, The energy function is expressed as: ； in, Indicates candidate outbound sequence Energy; This indicates the space utilization rate.

4. The order-driven mixed-box palletizing method according to claim 3, characterized in that, Two versions of space utilization assessment are used: Online version, This indicates the space utilization rate after simulating palletizing of the current pallet to be stacked; Offline version This represents the overall space utilization rate after performing a complete simulated palletizing of all goods in the order queue.

5. The order-driven mixed-box palletizing method according to claim 2, characterized in that, The hybrid optimization strategy employing a combination of local sampling using the integrated Langevin update operator and global evolution using a genetic algorithm to retrieve the optimal outbound sequence from candidate outbound sequences includes: A population is initialized based on the population size. The first individual in the population is the order queue, and the remaining individuals are randomly generated from the order queue. Each individual represents a candidate outbound sequence. Local optimization: For any candidate outbound sequence, the Langevin update operator generates a neighborhood solution by randomly swapping the outbound order of two goods in the sequence, and calculates the energy difference between the energy corresponding to the neighborhood solution and the energy corresponding to the original candidate outbound sequence; based on the energy difference, temperature parameters, and step size factor, the probability of an individual accepting a new solution is determined, and the individual state is updated according to this probability; Global evolution: After all candidate outgoing sequences have completed the local optimization of the Langevin update operator, the genetic algorithm performs the selection logic of survival of the fittest based on the energy value of the individual, selects the superior parent individuals from the current population for subsequent evolution, and the selected parent individuals generate offspring through the sequential crossover operator, and the offspring genes are fine-tuned by the mutation operator. The optimal outbound sequence is retrieved by iteratively optimizing the local and global processes described above.

6. The order-driven mixed-box palletizing method according to any one of claims 1-5, characterized in that, The online palletizing algorithm, based on the box's geometric dimensions and the three-dimensional state information of the current pallet to be palletized, plans the optimal placement position and orientation of the box on the current pallet to be palletized, and further includes: If the current waiting pallet is insufficient to accommodate the boxes, the current waiting pallet is stored in the completed pallet set and a new empty waiting pallet is created to continue the operation until the splitting requirement of the order is met.

7. An order-driven mixed-box palletizing system, characterized in that, To implement the mixed-box palletizing method according to any one of claims 1-6, comprising: The input module is configured to: input order tasks and generate corresponding order sets; The initialization module is configured to sort the order set according to the weight of the goods, generate an initial order queue, and initialize the set of pallets to be unpacked, pallets to be stacked, and completed pallets to an empty set. The dynamic outbound module is configured to: when the pallet to be unloaded is idle, use an energy-based outbound order optimization algorithm to select the best matching goods from the order queue, send an outbound instruction to guide the automated warehouse to transfer the box corresponding to the best matching goods to the pallet to be unloaded, and remove the outbound goods from the order queue; In addition, the online palletizing module is configured to: perceive the geometric dimensions of the boxes on the pallet to be unpacked in real time, and based on the geometric dimensions of the boxes and the three-dimensional status information of the current pallet to be stacked, plan the optimal placement position and orientation of the boxes on the current pallet to be stacked and send corresponding commands to the operation execution module.

8. The order-driven mixed-box palletizing system according to claim 7, characterized in that, The operation execution module includes: an automated warehouse, pallets to be unloaded, pallets to be stacked, and a handling unit connecting the various areas; The automated warehouse will release the corresponding goods and transfer them to the pallets to be unloaded according to the instructions sent by the dynamic outbound module. The handling unit, in the area to be unpacked, executes the command issued by the online palletizing module to grab the box and place it on the pallet to be stacked; Once the target boxes on the pallet to be unloaded have been removed or the order requirements have been met, the remaining boxes are transferred back to the warehouse by the handling unit, triggering a closed-loop feedback mechanism to enter the next round of decision-making.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the mixed-box palletizing method according to any one of claims 1-6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the mixed-box palletizing method according to any one of claims 1-6.

Images