Shuttle vehicle dispatching management method and device, computer equipment and storage medium

Abstract

The invention discloses a shuttle vehicle scheduling management method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a target task, and mapping the target task into a semantic feature vector; obtaining at least one four-way shuttle vehicle corresponding to the target task at present, and performing parameter characterization processing on the at least one four-way shuttle vehicle to obtain a parameter feature vector corresponding to the four-way shuttle vehicle; calculating a correlation score of the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and selecting a target shuttle vehicle for executing the target task from the four-way shuttle vehicles based on the correlation score; and performing path planning on the target shuttle vehicle by adopting a heuristic algorithm, and performing conflict detection and resolution on the target shuttle vehicle in combination with a preset conflict resolution strategy. The problems of path conflict, response lag and the like in multi-vehicle cooperation can be solved, and high efficiency and safety of multi-vehicle cooperation operation are achieved.

Description

Technical Field

[0001] This invention relates to the field of warehousing and logistics technology, and in particular to a shuttle dispatching and management method, device, computer equipment, and storage medium. Background Technology

[0002] Four-way shuttle systems, as an emerging high-density automated warehousing solution, are widely used in modern logistics centers due to their ability to move in four directions within a plane and achieve cross-level operations with the help of elevators, offering advantages such as high space utilization and flexible layout. However, with the expansion of system scale and the increase in task complexity, multi-vehicle collaborative scheduling faces severe challenges, mainly manifested in: 1) the risk of conflicts and deadlocks in multi-vehicle paths; 2) insufficient precision in matching vehicle resources with dynamically changing task requirements; and 3) rigid cross-level scheduling strategies, resulting in reduced efficiency.

[0003] Existing technologies have attempted to address these issues. For example, some solutions construct digital twin models to conduct time-window-based A* algorithm path planning and verification in virtual space, thereby avoiding conflicts with real-world equipment. Other solutions focus on real-time obstacle avoidance in path planning, such as pre-locking waypoints and detecting occupancy to dynamically adjust strategies (normal driving, waiting, or replanning), thus reducing conflicts and inefficient transport. Further research optimizes the overall system operation time by employing intelligent algorithms, such as ant colony-particle swarm optimization, to globally optimize the order of inbound and outbound tasks. Additionally, improved algorithms based on the Conflict Search (CBS) framework have been applied to plan conflict-free optimal paths for multiple vehicles.

[0004] However, existing technologies still have the following shortcomings: First, most scheduling methods select vehicles based on static rules (such as nearest distance), failing to fully consider the dynamic and deep correlation between real-time vehicle status (such as battery level and current location) and task semantic requirements (such as urgency and cargo characteristics). Second, regarding cross-layer scheduling triggering conditions, the rules are usually relatively fixed, lacking intelligent prediction of overall system load balancing and future task queues. Third, path planning, task allocation, and conflict resolution are often phased or loosely coupled, lacking an integrated, adaptive optimization framework from task understanding to vehicle selection and conflict-free path generation.

[0005] Therefore, there is an urgent need for a scheduling method that can achieve intelligent, dynamic, and accurate mapping between tasks and vehicles, and integrate simulation optimization and real-time replanning capabilities, in order to further improve the overall operational efficiency, stability, and adaptability of the four-way shuttle system. Summary of the Invention

[0006] This invention provides a shuttle scheduling management method, apparatus, computer equipment, and storage medium, aiming to improve the scheduling effect of shuttles.

[0007] In a first aspect, embodiments of the present invention provide a shuttle dispatching and management method, comprising: A three-dimensional coordinate array of storage compartments in the warehouse is predefined, and a relative coordinate system is established based on the three-dimensional coordinate array; Based on the relative coordinate system, the target task is obtained and mapped into a semantic feature vector; Obtain at least one four-way shuttle corresponding to the target task, and perform parameter feature processing on at least one four-way shuttle to obtain the parameter feature vector corresponding to the four-way shuttle; The correlation score between the semantic feature vector and the parameter feature vector is calculated through a dynamic memory collaborative interaction network, and a target shuttle for performing the target task is selected from the four-way shuttles based on the correlation score. A heuristic algorithm is used to plan the path for the target shuttle, and a pre-set conflict resolution strategy is used to detect and resolve conflicts for the target shuttle.

[0008] Secondly, embodiments of the present invention provide a shuttle dispatching and management device, comprising: The coordinate system establishment unit is used to predefine the three-dimensional coordinate array of the storage compartment and establish a relative coordinate system to map the physical storage location into a simulation node; The task mapping unit is used to obtain the target task based on the relative coordinate system and map the target task into a semantic feature vector; The feature processing unit is used to obtain at least one four-way shuttle currently corresponding to the target task, and to perform parameter feature processing on at least one four-way shuttle to obtain the parameter feature vector corresponding to the four-way shuttle. The target selection unit is used to calculate the correlation score between the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and select a target shuttle from the four-way shuttles to perform the target task based on the correlation score. The planning and resolution unit is used to plan the path of the target shuttle using a heuristic algorithm, and to detect and resolve conflicts of the target shuttle in combination with a preset conflict resolution strategy.

[0009] Thirdly, embodiments of the present invention provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the shuttle scheduling and management method as described in the first aspect.

[0010] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the shuttle scheduling and management method as described in the first aspect.

[0011] This invention provides a shuttle scheduling and management method, apparatus, computer equipment, and storage medium. The method includes: pre-defining a three-dimensional coordinate array of storage compartments in a warehouse and establishing a relative coordinate system based on the three-dimensional coordinate array; obtaining a target task based on the relative coordinate system and mapping the target task to a semantic feature vector; obtaining at least one four-way shuttle currently corresponding to the target task and performing parameter feature processing on the at least one four-way shuttle to obtain a parameter feature vector corresponding to the four-way shuttle; calculating the correlation score between the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and selecting a target shuttle from the four-way shuttles to execute the target task based on the correlation score; performing path planning for the target shuttle using a heuristic algorithm and combining a preset conflict resolution strategy to perform conflict detection and resolution for the target shuttle. This invention deeply associates the target task with the real-time status and semantic requirements of the four-way shuttle, and dynamically memorizes and coordinates these with the real-time operating parameters of the four-way shuttle. This enables intelligent, dynamic, and precise mapping between tasks and vehicles, thus solving problems such as path conflicts, response delays, and inefficient layer switching in multi-vehicle collaboration. It can significantly reduce the system conflict rate and improve task response speed, thereby achieving high efficiency and safety in multi-vehicle collaborative operations and improving the scheduling effect of the four-way shuttle. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart illustrating a shuttle dispatching and management method provided in an embodiment of the present invention; Figure 2 A schematic block diagram of a shuttle dispatching and management device provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of cargo location coordinate mapping in a shuttle dispatching and management method provided by an embodiment of the present invention; Figure 4 This is a schematic diagram of a simulation model in a shuttle dispatching and management method provided in an embodiment of the present invention; Figure 5 This is an example diagram of the overall scheduling in a shuttle dispatching and management method provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the path planning and conflict resolution logic in a shuttle scheduling and management method provided in an embodiment of the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0016] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0017] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0018] Please see below. Figure 1 The present invention provides a shuttle scheduling and management method, which specifically includes steps S101 to S105.

[0019] Step S101: Predefine the three-dimensional coordinate array of the storage compartments in the warehouse, and establish a relative coordinate system based on the three-dimensional coordinate array; Step S102: Based on the relative coordinate system, obtain the target task and map the target task into a semantic feature vector; Step S103: Obtain at least one four-way shuttle corresponding to the target task, and perform parameter feature processing on at least one four-way shuttle to obtain the parameter feature vector corresponding to the four-way shuttle. Step S104: Calculate the correlation score between the semantic feature vector and the parameter feature vector through the dynamic memory collaborative interaction network, and select the target shuttle from the four-way shuttles to perform the target task based on the correlation score. Step S105: Use a heuristic algorithm to plan the path for the target shuttle, and combine it with a preset conflict resolution strategy to detect and resolve conflicts for the target shuttle.

[0020] In this embodiment, firstly, a three-dimensional coordinate array is set for the warehouse storage grid, and a relative coordinate system is constructed accordingly to map the physical storage locations to the simulation nodes. Next, on the constructed relative coordinate system, the target task is received and converted into a semantic feature vector. Then, the four-way shuttle currently capable of executing the task is acquired, and its parameters are characterized to generate a parameter feature vector. Afterward, a dynamic memory collaborative interaction network is used to calculate the correlation score between the semantic feature vector and the parameter feature vector, and the target shuttle for executing the task is selected from the available shuttles based on the score. Finally, a heuristic algorithm is used to plan a path for the target shuttle, and a preset conflict resolution strategy is combined to detect and handle conflicts.

[0021] This embodiment deeply correlates the target task with the real-time status and semantic requirements of the four-way shuttle, fully considering the dynamic and deep correlation between the vehicle's real-time status (such as battery level and current location) and the task's semantic requirements (such as urgency and cargo characteristics). This deep correlation makes the selection of vehicles to perform tasks more accurate and reasonable, avoiding the limitations of selecting vehicles based on static rules. Simultaneously, by combining a dynamic memory collaborative interaction network to calculate relevance scores, vehicle selection can be adjusted in real time according to changes in the vehicle's real-time status and task semantic requirements, improving the flexibility and adaptability of task allocation. Regarding path planning and conflict resolution, a heuristic algorithm is used for path planning of the target shuttle, combined with a pre-set conflict resolution strategy for conflict detection and resolution, achieving integrated and adaptive optimization of path planning, task allocation, and conflict resolution. Unlike the phased or loosely coupled approach of existing technologies in path planning, task allocation, and conflict resolution, this embodiment comprehensively optimizes from task understanding to vehicle selection and then to conflict-free path generation, reducing path conflicts and ineffective handling in multi-vehicle collaboration and improving the overall operational efficiency of the system. In summary, this embodiment deeply associates the target task with the real-time status and semantic requirements of the four-way shuttle, and combines a dynamic memory collaborative interaction network and heuristic algorithms to achieve intelligent, dynamic, and accurate mapping between tasks and vehicles. It solves problems such as path conflicts, response lag, and inefficient layer switching in multi-vehicle collaboration, significantly reduces the system conflict rate, improves task response speed, and enhances the scheduling effect of the four-way shuttle. It provides a more efficient and intelligent solution for multi-vehicle collaborative scheduling in the warehousing and logistics field.

[0022] In one embodiment, the predefined three-dimensional coordinate array of storage compartments in the warehouse, and the establishment of a relative coordinate system based on the three-dimensional coordinate array, includes: The three-dimensional coordinate array is defined as follows: ε=(X i ,Y j Z k ); X i =A×i+α×C; Y j =B×j+β×D; Z k =E×a+F×b+...+Z'×n; Where ε represents the three-dimensional coordinate array, A is the length of the storage compartment in the X direction, i is the storage compartment number in the X direction, α is the width of the sub-channel in the X direction, C is the number of sub-channels in the X direction; B is the length of the storage compartment in the Y direction, j is the storage compartment number in the Y direction; β is the width of the sub-channel in the Y direction, D is the number of sub-channels in the Y direction; E, F, ..., Z' are different floor heights, a, b, ..., n are the corresponding floor numbers, and k = a + b + ... + n; Based on the three-dimensional coordinate array, a relative coordinate system is established with the center of the first storage cell in the lower left corner of the warehouse as the origin, so as to map the physical storage location as a simulation node.

[0023] In this embodiment, combined with Figures 3-5 For intensive storage warehouses, the grid coordinates are first defined as a three-dimensional array ε=(X i ,Y j Z k ): where i, j, and k represent the number of storage locations in the X (column), Y (row), and Z (layer) directions, respectively; the X coordinate corresponds to the position of the storage location in the X direction (i storage locations in this direction), the Y coordinate corresponds to the position of the storage location in the Y direction (j storage locations in this direction), and the Z coordinate is the height of the storage cell from the ground (corresponding to k layers of the warehouse structure). This three-dimensional coordinate system uniquely determines the location of any storage location. Subsequently, a relative coordinate system is established with the center of the first storage cell in the lower left corner as the origin, mapping the physical storage locations to simulation nodes. In this way, the system can be analyzed and optimized within the constructed simulation environment.

[0024] Here, for the shuttle scheduling and management problem, the accurate representation of the storage grid coordinates and the mapping from physical storage locations to simulation nodes are crucial foundational steps. By establishing the aforementioned three-dimensional coordinate array and relative coordinate system, the positional relationships of each storage location within the warehouse can be clearly described. This provides precise location information for subsequent operations such as task mapping, vehicle status assessment, and path planning. For example, during task mapping, based on the starting and target storage locations' three-dimensional coordinates, the task can be accurately transformed into a semantic feature vector. These semantic feature vectors can contain important information such as the task's starting position, target position, and task type, providing a basis for subsequently selecting a suitable four-way shuttle to execute the task.

[0025] In one embodiment, the step of calculating the correlation score between the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and selecting a target shuttle from the four-way shuttles to perform the target task based on the correlation score, includes: The correlation score between the semantic feature vector and the parameter feature vector is calculated using an attention mechanism; The correlation score is set as the parameter weight of the four-way shuttle. The dynamic score of each of the four-way shuttle vehicles is calculated based on the parameter weights. Based on the dynamic score, the four-way shuttle with the highest score is selected as the target shuttle.

[0026] In this embodiment, when selecting a target shuttle, an attention mechanism is first used to calculate the relevance score between the semantic feature vector and the parameter feature vector. This mechanism focuses on the parts of the feature vector that are highly relevant to the task, making the calculation results more accurately reflect the matching degree between the task and the vehicle. The obtained relevance score is set as the parameter weight of the four-way shuttle, meaning that when different vehicles perform tasks, their various parameters will be weighted to different degrees according to their matching degree with the current task. Then, the dynamic score of each four-way shuttle is calculated through the parameter weights, thus comprehensively considering the real-time status of the vehicle and the semantic requirements of the task. For example, for an emergency task, a vehicle with sufficient battery power and close to the task starting point may obtain a higher dynamic score due to its higher parameter weight. Subsequently, the target shuttle can be selected based on the dynamic score results, such as selecting the four-way shuttle with the highest dynamic score or one of the top four-way shuttles in terms of dynamic score. This selection method ensures that each task is assigned to the most suitable vehicle, improving the efficiency and quality of task execution. Compared to traditional vehicle selection methods based on static rules or simple matching, this embodiment, through a dynamic memory collaborative interaction network and attention mechanism, can more intelligently consider the real-time status of vehicles and the semantic requirements of tasks, achieving precise matching between tasks and vehicles. This effectively reduces conflicts and inefficient handling in multi-vehicle collaborative operations. For example, it avoids assigning tasks to vehicles that are too far away or have insufficient power, thereby reducing vehicle travel distance and waiting time, and improving the overall operational efficiency of the system. Furthermore, because it can adjust vehicle selection in real time according to changes in vehicle status and task semantic requirements, this embodiment also improves the flexibility and adaptability of task allocation, enabling the shuttle dispatch management system to better cope with various complex warehousing and logistics scenarios.

[0027] In practical applications, the first step is to create a physical model of the task and its environment. For example, in Flexsim, inbound and outbound orders are modeled as temporary entities called Items, and their labels store the coordinates (X, Y, F, G) of the target storage location. i ,Y j ) and floor Z k Information. The storage locations, track intersections, and elevator connection points are modeled as a node-based path network. Dynamic management of warehouse resource occupancy is achieved through node status control. Then, the multimodal parameters of the four-way shuttle are characterized. For example, three types of status labels are assigned to the four-way shuttle: real-time updated current battery level (CE), estimated travel distance to the target point (D), and expected number of reversing nodes (N). Task descriptions such as "emergency replenishment" are mapped to task type codes. In the simulation, a coding module is constructed using custom logic to normalize the three types of numerical parameters and obtain the battery level feature (E). ce Distance feature E d commutation characteristic E n The task type is encoded and mapped to a semantic feature vector E through a built-in semantic table.task Then, through a dynamic memory collaborative interaction network, E... task respectively with E ce E d E n The network simulates the cognitive process of "dynamically recalling and weighing the importance of various parameters according to task requirements." E is calculated through an attention mechanism. task The relevance score of each parameter feature is generated in real time as the dynamic weight ratio (W) of that parameter in the current task context. ce W d W n Using dynamically generated weight ratios, the dynamic score for each candidate shuttle can be calculated: ; Where f, g, and h are the standardized scoring functions for each parameter, such as high battery capacity, short distance, and few commutations, which result in high scores. ω, θ, These are adjustable base coefficients. Based on the dynamic score, the scheduler (or the system, server, or other implementing entity) can select the vehicle with the highest score and assign the task using the dispatchjob() function.

[0028] In one embodiment, the step of using a heuristic algorithm to plan the path for the target shuttle and combining it with a preset conflict resolution strategy to detect and resolve conflicts for the target shuttle includes: The shortest path to the target shuttle is calculated using the A* algorithm and the Manhattan distance heuristic function. The shortest path calculation result is set as the path planning result, and the target shuttle is controlled to execute the target task according to the path planning result.

[0029] Combination Figure 6 In this embodiment, the A* algorithm combined with the Manhattan distance heuristic function is used to calculate the shortest path for the target shuttle during path planning. The A* algorithm is a commonly used pathfinding algorithm that efficiently finds the shortest path by comprehensively considering both the actual cost from the starting point to the current point and the estimated cost from the current point to the target point. The Manhattan distance heuristic function provides the algorithm with reasonable heuristic information, helping it to search more directionally, reduce unnecessary search range, and improve the efficiency of path planning.

[0030] After obtaining the shortest path, it is used as the final solution for path planning. This ensures that the target shuttle can reach the target location with the shortest distance and in the least amount of time. Subsequently, the system controls the target shuttle to execute the target task according to the planned path, ensuring that the task can be completed quickly and accurately.

[0031] In a specific embodiment, the shortest path is calculated according to the following formula: h(n) = |x 目标 -x n |+|y 目标 -y n |; Among them, (x n ,y n (x) represents the coordinates of the current node. 目标 ,y 目标 () represents the coordinates of the target point.

[0032] In one embodiment, the step of using a heuristic algorithm to plan the path for the target shuttle and combining it with a preset conflict resolution strategy to detect and resolve conflicts for the target shuttle further includes: A dual-mode conflict detection mechanism is used to detect conflicts between the target shuttle and other four-way shuttles; wherein, the dual-mode conflict detection mechanism includes a same-track same-direction rear-end collision conflict mechanism and a different-track crossover conflict mechanism; When a conflict is detected between the target shuttle and other four-way shuttles, a conflict resolution strategy is adopted to schedule and control the target shuttle. The conflict resolution strategy includes: prioritizing the use of a waiting mechanism to control the passage order of the target shuttle and other four-way shuttles, and when the strategy of controlling the passage order cannot avoid conflict, controlling other four-way shuttles to wait in front of the corresponding conflict node until a safe passage time is reached.

[0033] Combination Figure 6 In this embodiment, when performing conflict detection and resolution, a dual-mode conflict detection mechanism is first employed: a same-track, same-direction rear-end collision conflict mechanism and a different-track intersection conflict mechanism. This mechanism comprehensively detects conflicts between the target shuttle and other shuttles traveling in the same direction. The same-track, same-direction rear-end collision conflict mechanism focuses on shuttles traveling in the same direction on the same track, determining the risk of a rear-end collision by monitoring vehicle speed, position, and spacing in real time. The different-track intersection conflict mechanism, on the other hand, focuses on shuttles at intersections of different tracks, analyzing their travel paths and timing to detect whether a collision will occur at the intersection.

[0034] When a conflict is detected between a target shuttle and other four-way shuttles, the target shuttle will be strictly controlled according to the conflict resolution strategy. A waiting mechanism is prioritized to control the passage order of the target shuttle and other four-way shuttles. This mechanism rationally arranges the waiting time and passage order of vehicles based on factors such as the urgency of the task and the direction of travel. For example, vehicles with urgent tasks can pass through the conflict area first, while other vehicles must wait. However, in some complex situations, relying solely on the strategy of controlling the passage order may not completely avoid conflicts. In this case, the system will control other four-way shuttles to wait at the corresponding conflict node until a safe passage time is reached. This measure ensures that no vehicle collisions occur at the conflict node, guaranteeing the safe operation of the shuttles.

[0035] By combining this dual-mode conflict detection mechanism and conflict resolution strategy, conflicts in multi-vehicle collaborative operations can be effectively reduced, improving the safety and stability of the shuttle dispatch management system. Simultaneously, this real-time monitoring and dynamic adjustment approach enables the system to better cope with various complex warehousing and logistics scenarios, further enhancing the overall operational efficiency. In practical applications, this conflict detection and resolution method can significantly reduce the risk of vehicle collisions and minimize operational delays caused by conflicts, providing strong support for the efficient operation of warehousing and logistics.

[0036] For example, in a rear-end collision on the same track in the same direction, if the time it takes for the preceding vehicle to reach the intersection is t1 and the time it takes for the following vehicle to reach the intersection is t2, and the safe time interval between the two vehicles is set to Δt, then if t2 ≥ t1 + Δt, a rear-end collision is considered to have occurred; otherwise, a rear-end collision is determined to have occurred. For a collision at an intersection on different tracks, assuming the two vehicles will meet at the intersection, the arrival times of the two vehicles at the intersection are calculated as t1 and t2 respectively. a and t b Let τ be the time it takes for the two cars to pass the node (which can be understood as the time required for turning). Then, when t... a With t bIf the time difference between the two points is greater than or equal to τ, it is considered that no cross-track conflict will occur; otherwise, a cross-track conflict is determined to exist. When a conflict is determined to exist, scheduling can be quickly carried out according to the conflict resolution strategy. If a waiting mechanism is adopted, the system will comprehensively consider factors such as task priority, vehicle remaining battery power, and distance to the target cargo location to determine which vehicle should wait first and for how long. For example, vehicles performing high-priority emergency replenishment tasks will be prioritized to pass through the conflict area, while other vehicles will have to wait for a suitable opportunity. In addition, while controlling other four-way shuttles to wait before the conflict node until the safe passage time, the system will monitor the vehicle's position and speed in real time and accurately calculate the safe passage time node. At the same time, when it is necessary to schedule another vehicle to delay its journey or take a detour, the system can calculate the delay time (safe interval time - current time difference) and the detour time (alternative path length - original path length) to select the solution with the shorter total time.

[0037] To further improve the system's adaptability and flexibility, the conflict detection and resolution strategies can be dynamically adjusted. For example, the sensitivity of the dual-mode conflict detection mechanism can be adjusted based on the volume of warehousing and logistics operations during different time periods. During peak business hours, the detection frequency can be increased to detect potential conflicts earlier; during relatively quiet business hours, the detection frequency can be appropriately reduced to decrease system resource consumption.

[0038] In one embodiment, the shuttle dispatching and management method further includes: The task queues and vehicle distribution at each layer of the warehouse are monitored globally to determine whether a layer-switching scheduling mechanism is triggered. The triggering conditions for the layer-switching scheduling mechanism are: the number of tasks in the current layer is less than the corresponding number of vehicles, the backlog of tasks in other layers is greater than a preset backlog threshold, and the total cross-layer scheduling distance of the four-way shuttle is less than the equivalent distance to return to the current layer for standby. If the layer-switching scheduling mechanism is triggered, then the four-way shuttle cars in each layer of the warehouse will be switched between layers.

[0039] In addition to employing conflict resolution strategies to improve the scheduling efficiency of the four-way shuttle, this embodiment also utilizes a layer-switching scheduling mechanism to further optimize system resource allocation. By globally monitoring the task queues and vehicle distribution at each layer of the warehouse, the status of tasks and vehicles at each layer can be promptly grasped. This information is then used to determine whether the triggering conditions for the layer-switching scheduling mechanism are met. Here, for the triggering conditions of the layer-switching scheduling mechanism, if the number of tasks in the current layer is less than the corresponding number of vehicles, it means that the vehicle resources in the current layer are relatively abundant; while the backlog of tasks in other layers exceeds a preset backlog threshold, it indicates that there are many tasks waiting to be processed in other layers; at the same time, the total cross-layer scheduling distance of the four-way shuttle is less than the equivalent distance to return to the current layer for standby, which ensures that the cross-layer scheduling is reasonable in terms of distance and does not cause excessive resource waste.

[0040] Once the layer-swapping scheduling mechanism is triggered, the four-way shuttles on each level of the warehouse are reassigned. During this process, multiple factors are considered. For example, vehicles are allocated based on the urgency and type of tasks on each level. For urgent outbound tasks, vehicles can be prioritized for dispatch to the corresponding level. Additionally, vehicle battery levels are considered; if a vehicle has low battery, it may not be suitable for cross-level scheduling to avoid insufficient power during the scheduling process.

[0041] In practice, by monitoring the task queues and vehicle status information at each layer in real time, algorithms can automatically determine whether to trigger the layer-switching scheduling mechanism. Once triggered, a layer-switching scheduling instruction is generated, guiding the corresponding four-way shuttle to the target layer. During the vehicle layer-switching process, it is also necessary to combine this with the previously mentioned conflict detection and resolution strategies to ensure that vehicles do not conflict with other vehicles during cross-layer movement. This layer-switching scheduling mechanism allows for a more rational allocation of vehicle resources across warehouse layers, improving the overall operational efficiency of the system and achieving global load balancing. It avoids situations where vehicles on some layers are idle while tasks accumulate on others, ensuring that tasks at each layer are processed promptly. Simultaneously, the combination of conflict detection and resolution strategies guarantees the safe operation of vehicles during layer-switching, further enhancing the performance and adaptability of the shuttle dispatch management system, enabling it to better handle different warehousing and logistics scenarios.

[0042] In one embodiment, the shuttle dispatching and management method further includes: The operating status of the target shuttle is monitored in real time, and it is determined whether a locking mechanism is triggered; wherein, the triggering condition of the locking mechanism includes the target shuttle waiting time at the corresponding node exceeding a preset time threshold. If the locking mechanism is triggered, a new target shuttle will be selected.

[0043] This embodiment further ensures the stable operation of the shuttle dispatch management system by monitoring the operational status of the target shuttle in real time. When the waiting time of the target shuttle at the corresponding node exceeds a preset time threshold, it indicates that the vehicle may have encountered an abnormal situation, such as a malfunction or being blocked by other vehicles. At this time, a locking mechanism can be triggered. After the locking mechanism is triggered, the original plan can be interrupted, that is, the target shuttle is controlled to stop executing the target task, and a new target shuttle is selected to execute the target task based on the latest system status (such as the status parameters of the current four-way shuttles in the warehouse). The process of reselecting the target shuttle is consistent with the previous vehicle selection method. It still determines the most suitable vehicle by calculating a dynamic score based on the changes in the real-time status of the vehicle and the semantic requirements of the task, thereby realizing the system's self-recovery and continuous optimization.

[0044] This locking mechanism enables timely detection and handling of anomalies in target shuttles, preventing task delays caused by vehicle malfunctions or other reasons. Simultaneously, selecting a new target shuttle ensures smooth task execution, improving the reliability and stability of the shuttle dispatch management system. Furthermore, combined with existing mechanisms for flexible task allocation, path planning, conflict detection and resolution, and layer-by-layer scheduling, a complete, efficient, and intelligent shuttle dispatch management system is formed. This system better addresses various complex and changing warehousing and logistics scenarios, improving operational efficiency and service quality across the entire warehousing and logistics process, saving costs and creating more value for enterprises.

[0045] The following is a specific example to illustrate the shuttle dispatching and management method described in this embodiment. In this example, the length of the storage compartments in the X direction is A = 1.2m, and the length in the Y direction is B = 0.8m; the width of the sub-aisles in the X direction is α = 1.2m, and the number of sub-aisles is C = 4; the width of the main aisle in the Y direction is β = 2.0m, and the number of main aisles is D = 2; the height specifications of the storage compartments include E = 2.0m (a = 1 compartment) and F = 2.2m (b = 2 compartments). Therefore, the number of storage compartment layers is k = 1 + 2 = 3 layers, and the height of each layer is as follows: Z1 = 2.0 × 1 = 2.0 m; Z2 = 2.0 × 1 + 2.2 × 1 = 4.2 m; Z3 = 2.0 × 1 + 2.2 × 2 = 6.4 m; First, the coordinates of the cargo location are mapped to the track nodes. The X-direction coordinates are then calculated using the equation: X i =1.2×i+1.2×4=1.2i+4.8, i=1,2,…; The Y-axis coordinate is: Y j =0.8×j+2.0×2=0.8j+4.0, j=1,2,…; This forms the grid coordinate array: ε=(1.2i+4.8,0.8j+4.0,2.0)+(1.2i+4.8,0.8j+4.0,4.2)+(1.2i+4.8,0.8j+4.0,6.4); Path planning: Load the A* module in Flexsim and set up a four-way shuttle to travel from node (6.0, 4.8, 4.2) to the heuristic function (22.8, 19.2, 4.2). The shortest path is: h(n) = x 目标 -x n +y 目标 -y n =|22.8−6|+|19.2−4.2|=31.8m; A safe interval time Δt = 3s is set, and the turning time τ = 5s. If a vehicle traveling in the same direction on the same track arrives at node v at t1 = 10s, the following vehicle must satisfy t2 ≥ 13s. If the arrival times of the two vehicles at intersection node u are 15s and 17s respectively, with an interval of 2s < 5s, then one vehicle's arrival time is adjusted to be delayed by 3s. Simultaneously, the main passage is used for rapid passage and lane changing, while the sub-passage is used for receiving and releasing goods. Empty vehicles pass under the racks, while fully loaded vehicles detour.

[0046] Then, dynamic task mapping for the four-way shuttle is performed: Assuming there are 3 shuttles (V1, V2, V3) in the system, a new task arrives: "Emergency replenishment" to the storage location (6.0, 4.8, 4.2).

[0047] Step ① Task materialization: The arriving new task is created as a temporary entity Item, whose label stores the target coordinates (22.8, 19.2, 4.2) and the task type code TE.

[0048] Step ② Parameter Feature Encoding: The system reads the real-time status of the three vehicles and performs normalization processing. Details are as follows: Vehicle V1: Battery CE is 95%, after conversion E ce1 The value is 0.95; the distance D to the target is 15 meters, and after conversion, E d1 The value is set to 0.3 (assuming the maximum distance is 50 meters, the normalized calculation would be 1-15 / 50=0.7, but considering that the g(D) function may have reciprocals or negative correlations, E is set to 0.3 for simplification). d The smaller the value, the closer the distance, hence E d1 =15 / 50=0.3); the expected number of reversals N is 4, and after conversion E n1 The value is 0.4; Vehicle V2: Battery level CE is 60%, distance D is 8 meters, estimated number of reversals N is 2, and after conversion, Ec e2 The value is 0.60, E d2 The value is 0.16, E n2 The value is 0.2; Vehicle V3: Battery level CE is 80%, distance D is 25 meters, estimated number of reversals N is 1, after conversion E ce3 The value is 0.80, E d3 The value is 0.50, E n3 The value is 0.1.

[0049] Task type TE is mapped to feature vector E through a semantic table. task=[0.1,0.8,0.1] (Assuming this encoding represents the following meanings: low power requirement, weight 0.1; high distance sensitivity, weight 0.8; low requirement for commutation number, weight 0.1).

[0050] Step ③ Dynamic weight generation: Calculate E using an attention mechanism task The correlation between the parameters and features. Calculations show that the correlation with the distance feature E... d The correlation was highest (because E) task The weight for mid-range distance is 0.8 (maximum). Assume the generated dynamic weight ratio is: W ce =0.1, W d =0.8, W n =0.1. This indicates that in the current "emergency replenishment" task scenario, the system is more concerned with the distance between the vehicle and the target.

[0051] Step 4: Dynamic score calculation: Set the base coefficient ω=θ= =1.0, the scoring function f(CE)=CE, g(D)=1-D / 50, h(N)=1-N / 10, then: V1 score: Score1 = 0.1 × 0.95 + 0.8 × (1 - 0.3) + 0.1 × (1 - 0.4) = 0.095 + 0.56 + 0.06 = 0.715; V2 score: Score2 = 0.1 × 0.60 + 0.8 × (1 - 0.16) + 0.1 × (1 - 0.2) = 0.06 + 0.672 + 0.08 = 0.812; V3 score: Score3=0.1×0.80+0.8×(1-0.50)+0.1×(1-0.1)=0.08+0.40+0.09=0.570.

[0052] Therefore, V2 scored the highest. Although its battery level wasn't the highest, it best met the requirements of the "urgent" task because it was closest to the target and had fewer directional changes. Thus, the task was assigned to V2 using the dispatchJob() function.

[0053] Step 5: Global Collaborative Layer Change Decision (Extended to Multi-Layer Cases): Assume there is a three-layer warehouse. The global table presents the following information: Level 1: Number of missions: 3; Number of vehicles: 2. Tier 2: 10 missions, 3 vehicles; Tier 3: 1 mission, 2 vehicles.

[0054] According to the rule "the number of tasks in the current layer is less than the number of vehicles and the backlog of tasks in other layers is greater than 5", whether the first layer meets the condition needs to be determined in conjunction with the specific rules (although 3 < 2 is not true, let's assume the rule is that the number of tasks is less than the number of vehicles and the number of tasks in other layers is greater than 5). The second layer has a backlog of 10 tasks, which is greater than 5, and there are idle vehicles in the first layer. Therefore, the system may trigger a scheduling operation, using the elevator to dispatch an idle vehicle from the first layer to the second layer to perform a task, thereby achieving load balancing.

[0055] When vehicle V2 performs the aforementioned emergency task, while en route to the target node, a trigger detects that the node will be occupied for an extended period due to a vehicle malfunction ahead (BLOCK mechanism). First, a monitoring and alarm process is initiated. If V2 waits for more than 15 seconds at the node preceding it on the predetermined path, the system immediately triggers a deadlock warning. Then, path replanning is performed: the central scheduler interrupts V2's original path plan and, based on the latest global reservation table that marks the malfunctioning node and affected areas, uses V2's current position as the new starting point, maintaining the target, and re-executes dynamic decision-making and path planning. For the replanned new path, the improved A* algorithm avoids the malfunctioning area, planning a detour. Although the path length increases, the continuous execution of the task is ensured, achieving system self-recovery and continuous optimization.

[0056] In summary, this embodiment has the following advantages compared to the prior art: (1) Path efficiency optimization: By combining the Manhattan distance heuristic function with the A* algorithm, the length of the global path is reduced; (2) Reduced collision rate: By employing a dual collision detection mechanism and a waiting / detour strategy, the risk of collisions is reduced; (3) Improved resource utilization: Dynamic task mapping and cross-layer scheduling strategies are adopted to reduce idle driving and invalid waiting. (4) Modularization and enhanced practicality: By utilizing the built-in code, global table and triggers of Flexsim, core modules such as parameter encoding and interactive network can be flexibly built and adjusted to enhance the robustness of the system.

[0057] Figure 2 This is a schematic block diagram of a shuttle dispatching and management device 200 provided in an embodiment of the present invention. The device 200 includes: The coordinate system establishment unit 201 is used to predefine the three-dimensional coordinate array of the storage grid and establish a relative coordinate system to map the physical storage location into a simulation node; The task mapping unit 202 is used to obtain the target task based on the relative coordinate system and map the target task into a semantic feature vector; The feature processing unit 203 is used to obtain at least one four-way shuttle currently corresponding to the target task, and to perform parameter feature processing on at least one four-way shuttle to obtain the parameter feature vector corresponding to the four-way shuttle. The target selection unit 204 is used to calculate the correlation score between the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and select a target shuttle from the four-way shuttles to perform the target task based on the correlation score. The planning and resolution unit 205 is used to perform path planning for the target shuttle using a heuristic algorithm, and to perform conflict detection and resolution for the target shuttle in combination with a preset conflict resolution strategy.

[0058] In one embodiment, the coordinate system establishment unit 201 includes: An array definition unit is used to define the three-dimensional coordinate array according to the following formula: ε=(X i ,Y j Z k ); X i =A×i+α×C; Y j =B×j+β×D; Z k =E×a+F×b+...+Z'×n; Where ε represents the three-dimensional coordinate array, A is the length of the storage compartment in the X direction, i is the storage compartment number in the X direction, α is the width of the sub-channel in the X direction, C is the number of sub-channels in the X direction; B is the length of the storage compartment in the Y direction, j is the storage compartment number in the Y direction; β is the width of the sub-channel in the Y direction, D is the number of sub-channels in the Y direction; E, F, ..., Z' are different floor heights, a, b, ..., n are the corresponding floor numbers, and k = a + b + ... + n; The location mapping unit is used to establish the relative coordinate system based on the three-dimensional coordinate array, with the center of the first storage cell in the lower left corner of the warehouse as the origin, so as to map the physical location as a simulation node.

[0059] In one embodiment, the target selection unit 204 includes: A related calculation unit is used to calculate the correlation score between the semantic feature vector and the parameter feature vector through an attention mechanism; A weight setting unit is used to set the correlation score as the parameter weight of the four-way shuttle. The scoring calculation unit is used to calculate the dynamic score of each of the four-way shuttles based on the parameter weights. The scoring selection unit is used to select the four-way shuttle with the highest score as the target shuttle based on the results of dynamic scoring.

[0060] In one embodiment, the planning resolution unit 205 includes: The path calculation unit is used to calculate the shortest path for the target shuttle using the A* algorithm and the Manhattan distance heuristic function. The planning and execution unit is used to set the shortest path calculation result as the path planning result and control the target shuttle to execute the target task according to the path planning result.

[0061] In one embodiment, the planning resolution unit 205 further includes: The collision detection unit is used to perform collision detection between the target shuttle and other four-way shuttles using a dual-mode collision detection mechanism; wherein, the dual-mode collision detection mechanism includes a same-track same-direction rear-end collision mechanism and a different-track crossover collision mechanism. The conflict resolution unit is used to schedule and control the target shuttle by adopting a conflict resolution strategy when a conflict is detected between the target shuttle and other four-way shuttles. The conflict resolution strategy includes: prioritizing the use of a waiting mechanism to control the passage order of the target shuttle and other four-way shuttles, and controlling other four-way shuttles to wait in front of the corresponding conflict node until a safe passage time is reached when the strategy of controlling the passage order cannot avoid conflict.

[0062] In one embodiment, the shuttle dispatching and management device 200 further includes: The monitoring and judgment unit is used to perform global monitoring of the task queues and vehicle distribution of each layer of the warehouse to determine whether the layer switching scheduling mechanism is triggered. The triggering conditions of the layer switching scheduling mechanism are: the number of tasks in the current layer is less than the corresponding number of vehicles, and the backlog of tasks in other layers is greater than a preset backlog threshold. At the same time, the total cross-layer scheduling distance of the four-way shuttle is less than the equivalent distance to return to the current layer for standby. The layer-switching scheduling unit is used to perform layer-switching scheduling on the four-way shuttle cars in each layer of the warehouse if it is determined that the layer-switching scheduling mechanism has been triggered.

[0063] In one embodiment, the shuttle dispatching and management device 200 further includes: The locking judgment unit is used to monitor the operating status of the target shuttle in real time and determine whether the locking mechanism is triggered; wherein, the triggering condition of the locking mechanism includes the target shuttle waiting time at the corresponding node exceeding a preset time threshold. The lock determination unit is used to select a new target shuttle if the lock mechanism is triggered.

[0064] Since the embodiments of the apparatus and the embodiments of the method correspond to each other, please refer to the description of the embodiments of the method for the embodiments of the apparatus, which will not be repeated here.

[0065] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed, can perform the steps provided in the above embodiments. The storage medium may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0066] This invention also provides a computer device, which may include a memory and a processor. The memory stores a computer program, and when the processor calls the computer program in the memory, it can implement the steps provided in the above embodiments. Of course, the computer device may also include various network interfaces, power supplies, and other components.

[0067] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of the claims of this application.

[0068] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A method for managing the dispatch of a shuttle vehicle, characterized by, The method comprises the following steps: predefining a three-dimensional coordinate array of the storage bins in the warehouse, and establishing a relative coordinate system according to the three-dimensional coordinate array; based on the relative coordinate system, obtaining a target task and mapping the target task into a semantic feature vector; obtaining at least one four-way shuttle vehicle corresponding to the target task at present, and performing parameter feature processing on the at least one four-way shuttle vehicle to obtain a parameter feature vector corresponding to the four-way shuttle vehicle; calculating the correlation score of the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and selecting a target shuttle vehicle for executing the target task from the four-way shuttle vehicles based on the correlation score; adopting a heuristic algorithm to perform path planning on the target shuttle vehicle, and combining a preset conflict resolution strategy to perform conflict detection and resolution on the target shuttle vehicle.

2. The method of claim 1, wherein, The method of predefining a three-dimensional coordinate array of the storage bins in the warehouse, and establishing a relative coordinate system according to the three-dimensional coordinate array, comprises: defining the three-dimensional coordinate array according to the following formula: ε = (X i ,Y j ,Z k ) ; X i = A x i + a x C; Y j = B x j + β x D; Z k =E×a+F×b+...+Z’×n; wherein ε represents the three-dimensional coordinate array, A is the length of the X-direction storage bin, i is the X-direction storage bin number, α is the width of the X-direction sub-channel, and C is the number of X-direction sub-channels; B is the length of the Y-direction storage bin, j is the Y-direction storage bin number; β is the width of the Y-direction sub-channel, and D is the number of Y-direction sub-channels; E, F, …, Z' are different layer heights, a, b, …, n are corresponding layer numbers, and k = a + b + … + n; based on the three-dimensional coordinate array, taking the center of the first storage bin at the lower left corner of the warehouse as the origin to establish the relative coordinate system, so as to map the physical storage location into a simulation node.

3. The method of claim 1, wherein, The method of calculating the correlation score of the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network, and selecting a target shuttle vehicle for executing the target task from the four-way shuttle vehicles based on the correlation score, comprises: calculating the correlation score of the semantic feature vector and the parameter feature vector through an attention mechanism; setting the correlation score as the parameter weight of the four-way shuttle vehicle; calculating the dynamic score of each four-way shuttle vehicle based on the parameter weight; according to the result of the dynamic score, selecting the four-way shuttle vehicle with the highest score as the target shuttle vehicle.

4. The method of claim 1, wherein, The method of adopting a heuristic algorithm to perform path planning on the target shuttle vehicle, and combining a preset conflict resolution strategy to perform conflict detection and resolution on the target shuttle vehicle, comprises: adopting an A-star algorithm and a Manhattan distance heuristic function to calculate the shortest path of the target shuttle vehicle; setting the calculation result of the shortest path as the result of path planning, and controlling the target shuttle vehicle to execute the target task according to the result of path planning.

5. The method of claim 1, wherein, The method of adopting a heuristic algorithm to perform path planning on the target shuttle vehicle, and combining a preset conflict resolution strategy to perform conflict detection and resolution on the target shuttle vehicle, further comprises: adopting a double-mode conflict detection mechanism to perform conflict detection on the target shuttle vehicle and other four-way shuttle vehicles; wherein the double-mode conflict detection mechanism comprises a same-track same-direction rear-end conflict mechanism and a different-track intersection conflict mechanism; When it is detected that the target shuttle vehicle and other four-way shuttle vehicles have a conflict, a conflict resolution strategy is used to schedule and control the target shuttle vehicle; wherein the conflict resolution strategy includes: preferentially using a waiting mechanism to control the passing order of the target shuttle vehicle and other four-way shuttle vehicles, and when the strategy of controlling the passing order cannot avoid the conflict, controlling other four-way shuttle vehicles to wait in front of the corresponding conflict node until a safe passing time.

6. The method of claim 1, wherein, Also includes: The task queue and vehicle distribution of each layer of the warehouse library are globally monitored to determine whether to trigger a layer changing scheduling mechanism; wherein the triggering condition of the layer changing scheduling mechanism is that the current layer task number is less than the corresponding vehicle number, and the task backlog of other layers is greater than a preset backlog threshold, and the total distance of the four-way shuttle vehicle cross-layer scheduling is less than the equivalent distance of returning to the standby layer; If it is determined to trigger the layer changing scheduling mechanism, the four-way shuttle vehicles in each layer of the warehouse library are scheduled to change layers.

7. The method of claim 1, wherein, Also includes: The running state of the target shuttle vehicle is monitored in real time, and it is determined whether to trigger a locking mechanism; wherein the triggering condition of the locking mechanism includes that the waiting time of the target shuttle vehicle at the corresponding node exceeds a preset time threshold; If it is determined to trigger the locking mechanism, a new target shuttle vehicle is selected.

8. A shuttle vehicle dispatch management apparatus characterized by comprising: Comprise: A coordinate system establishing unit is configured to predefine a three-dimensional coordinate array of a cargo grid and establish a relative coordinate system to map a physical storage location to a simulation node; A task mapping unit is configured to obtain a target task based on the relative coordinate system and map the target task to a semantic feature vector; A feature processing unit is configured to obtain at least one four-way shuttle vehicle corresponding to the target task at present and perform parameter feature processing on the at least one four-way shuttle vehicle to obtain a parameter feature vector corresponding to the four-way shuttle vehicle; A target selecting unit is configured to calculate a correlation score of the semantic feature vector and the parameter feature vector through a dynamic memory collaborative interaction network and select a target shuttle vehicle for executing the target task from the four-way shuttle vehicles based on the correlation score; A planning resolution unit is configured to use a heuristic algorithm to plan a path for the target shuttle vehicle and combine a preset conflict resolution strategy to detect and resolve a conflict of the target shuttle vehicle.

9. A computer device, comprising: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the shuttle vehicle scheduling management method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the shuttle vehicle scheduling management method of any one of claims 1 to 7.

Images