Warehouse logistics scheduling method and system based on cloud computing

Abstract

The invention discloses a warehouse logistics scheduling method and system based on cloud computing, and relates to the field of warehouse logistics management, and the method comprises the steps: collecting to-be-processed cargo information and available robot state information in a warehouse management system in real time based on a cloud scheduling platform; carrying out the clustering of the to-be-processed goods, carrying out the calculation and evaluation of a clustering result, and obtaining an optimal task partition division scheme; obtaining the shortest collision-free path from each available robot to each task partition, and calculating the path length of the shortest collision-free path; calculating path suitability, load suitability and electric quantity sufficiency between each available robot and each task partition; determining an optimal allocation scheme between the available robots and the task partitions based on the three features; and scheduling the available robots based on the optimal allocation scheme and monitoring the operation states of the robots in real time. According to the invention, real-time, efficient and balanced global integrated scheduling can be carried out on logistics robots and order tasks, and the collaborative operation capability of a robot cluster is improved.

Description

Technical Field

[0001] This invention relates to the field of warehousing and logistics management technology, and in particular to a cloud computing-based warehousing and logistics scheduling method and system. Background Technology

[0002] With the rapid development of e-commerce and intelligent manufacturing, the management methods of modern warehousing and logistics centers are evolving towards intelligence and efficiency. Among these advancements, multi-robot systems have become core equipment for improving warehouse sorting and handling efficiency. However, the key to improving warehouse operational efficiency lies in how to efficiently and in real-time schedule robot clusters to collaboratively complete massive order tasks.

[0003] Existing warehouse management systems, such as AGV / AMR intelligent warehouse sorting and handling systems, have many limitations in their warehouse robot scheduling technology. For example, existing warehouse robot scheduling technologies often use simple greedy algorithms for task allocation, failing to optimize from a global perspective, resulting in the system's total cost being difficult to optimize. At the same time, existing warehouse robot scheduling technologies usually separate task partitioning from robot scheduling operations into two independent problems. When partitioning, the real-time status and dynamic performance of the robots are not considered, resulting in a mismatch between the partitioning scheme and the actual execution capacity. Moreover, existing warehouse robot scheduling systems have poor scalability, and centralized scheduling algorithms have high computational complexity, making it difficult to cope with the real-time scheduling needs of large-scale robots and order volumes, and unable to utilize cloud-based elastic computing resources. Summary of the Invention

[0004] To achieve real-time, efficient, and balanced global integrated scheduling of logistics robots and order tasks in complex and dynamic warehousing environments, and to address the shortcomings of existing warehousing robot scheduling technologies that struggle to meet the real-time scheduling demands of large-scale robots and order volumes, and cannot leverage cloud-based elastic computing resources, this invention provides a cloud-based warehousing logistics scheduling method and system. The technical solution is as follows: In a first aspect, the present invention provides a cloud-based warehouse logistics scheduling method, the steps of which include: real-time collection of information on goods to be processed and the status information of available robots in the warehouse management system based on a cloud scheduling platform; clustering the goods to be processed based on the information on goods to be processed and calculating and evaluating the clustering results to obtain an optimal task partitioning scheme; generating the shortest collision-free path from each available robot to each task partition based on the optimal task partitioning scheme and a path planning algorithm, and calculating its path length; calculating the path adaptability, load adaptability, and power sufficiency between each available robot and each task partition based on the path length, the information on goods to be processed, and the status information of available robots; calculating the matching degree between each available robot and each task partition based on the three characteristics, and determining the optimal allocation scheme between available robots and task partitions; issuing task instructions based on the optimal allocation scheme, scheduling available robots, and monitoring their operation status in real time.

[0005] Preferably, a data collection cycle is set, and the cloud scheduling platform periodically collects goods order information and robot cluster status information from the warehouse management system through the Internet of Things interface in real time. The goods order information includes the quantity of goods to be processed, the three-dimensional spatial coordinate data and weight data of each goods to be processed; the robot cluster status information includes the current position coordinates, remaining power and current load capacity of each robot. The information data of idle robots in the robot cluster status information is extracted and the data of robots whose remaining power does not exceed the threshold are removed to obtain the number of available robots, as well as the two-dimensional position coordinate data, remaining power data and current load capacity data of each available robot.

[0006] Preferably, based on the number of available robots, a range of values ​​for the number of task partitions is set, with the number of task partitions not exceeding the number of available robots. The K-Means++ algorithm is used to sequentially select the number of task partitions within the set range as the number of clusters. The coordinate data of the goods to be processed are clustered, and the number of clusters obtained is the same as the number of selected task partitions. Each cluster corresponds to one task partition. Based on the selected number of task partitions and the corresponding clustering results, the mean of the coordinates of all goods to be processed within each task partition is used as the center coordinate of the corresponding task partition. The squared Euclidean distance between each goods to be processed within the task partition and its center coordinate is calculated. Based on the number of goods to be processed within the task partition, the squared Euclidean distances corresponding to all goods to be processed within the task partition are accumulated as the intra-cluster spatial feature of the corresponding task partition. Based on the selected number of task partitions, the intra-cluster spatial features of all task partitions are accumulated as the intra-cluster spatial compactness corresponding to the set value of the number of task partitions. Similarly, the intra-cluster spatial compactness corresponding to each value within the set range of the number of task partitions is obtained.

[0007] Preferably, based on the selected number of task partitions and the corresponding clustering results, the weight data of all goods to be processed in each task partition is extracted, the total weight of all goods to be processed in each task partition is calculated, the mean of the total weight of all task partitions is taken as the average load, and the square of the difference between the total weight and the average load of each task partition is taken as the inter-cluster load feature of the corresponding task partition. Based on the selected number of task partitions, the inter-cluster load features of all task partitions are accumulated and the ratio between the accumulated value and the number of task partitions is calculated. The square root of the ratio is taken as the inter-cluster load balance corresponding to the value of the number of task partitions. Similarly, the inter-cluster load balance corresponding to each value within the set range of the number of task partitions is obtained.

[0008] Preferably, the intra-cluster spatial compactness and inter-cluster load balancing for each value within the set range of the number of task partitions are normalized. Weighting coefficients for intra-cluster spatial compactness and inter-cluster load balancing are set based on the actual application scenario. For a given number of task partitions, the sum of the products of the normalized intra-cluster spatial compactness and inter-cluster load balancing for each value and their respective weighting coefficients is used as the comprehensive evaluation value for that number of task partitions. Similarly, the comprehensive evaluation value for each value within the set range of the number of task partitions is obtained. The number of task partitions corresponding to the minimum value among all comprehensive evaluation values ​​is taken as the optimal number of task partitions for adapting to the robot scale, and the clustering result corresponding to the optimal number of task partitions is taken as the optimal task partitioning scheme.

[0009] Preferably, based on the actual warehouse layout, a gridded map containing obstacle information is constructed. The size of each grid cell in the map is set according to the scene's accuracy requirements. Areas that the robot can pass through are marked as passable grids and assigned a value of 1, while areas that the robot cannot pass through are marked as obstacle grids and assigned a value of 1. Based on the optimal task partitioning scheme, the average of the planar coordinates of all unprocessed goods within a task partition is used as the planar center coordinates of the corresponding task partition. The planar center coordinates of each task partition and the current position coordinates of each available robot are obtained. The current position coordinates of each available robot are used as the path start point, and the center coordinates of each task partition are used as the path end point. The path planning algorithm generates the shortest path from the starting point to the end point of each path, which serves as the shortest collision-free path from the corresponding available robot to each task partition. Based on the gridded map, the sequence of passable grids for each path is obtained. Based on the set grid cell size, the product of the number of passable grids for each path and the grid cell size is used as the path length of the corresponding path, thus obtaining the path length from each available robot to each task partition.

[0010] Preferably, based on the path length from each available robot to each task partition, for a certain task partition, the maximum and minimum values ​​of the path lengths from all available robots to that task partition are extracted, the difference between the maximum and minimum values ​​is taken as the path baseline difference for that task partition, the difference between the path length from any available robot to that task partition and the minimum value is taken as the path difference for the corresponding available robot, the ratio of the path difference to the path baseline difference is calculated, and the result of subtracting the ratio from 1 is taken as the path fit between the corresponding available robot and the task partition. Similarly, the path fit between each available robot and each task partition is obtained. Based on the information of goods to be processed and the status information of available robots, the weight data of all goods to be processed and the current load capacity data of each available robot are extracted. Based on the optimal task partitioning scheme, the maximum value of the weight data of goods to be processed in each task partition is extracted as the maximum load corresponding to each task partition. When the maximum load of a task partition is less than or equal to the current load capacity of an available robot, the ratio between the maximum load and the current load capacity is used as the load adaptability between the corresponding available robot and the task partition. When the maximum load of a task partition is greater than the current load capacity of an available robot, the load adaptability between the corresponding available robot and the task partition is set to 0. Similarly, the load adaptability between each available robot and each task partition is obtained. Based on the status information of available robots, the remaining power data of each available robot is extracted to obtain the power consumption coefficient per unit distance and the basic power consumption of the robot in operation. Based on the path length from each available robot to each task partition, the product of the power consumption coefficient and the path length is added to the basic power consumption to obtain the theoretical power consumption of the corresponding path. Similarly, the theoretical power consumption from each available robot to each task partition is obtained. For a certain task partition, when the remaining power data of the available robot is greater than the theoretical power consumption from the available robot to the task partition, the difference between the remaining power and the theoretical power consumption is calculated. The ratio between the difference and the robot's rated full power is used as the power adequacy between the corresponding available robot and the task partition. When the remaining power data of the available robot is less than or equal to the theoretical power consumption from the available robot to the task partition, the power adequacy between the corresponding available robot and the task partition is set to 0. Similarly, the power adequacy between each available robot and each task partition is obtained.

[0011] Preferably, based on the path adaptability, load adaptability, and power sufficiency between each available robot and each task partition, the matching degree between each available robot and each task partition is obtained. A bipartite graph and weight matrix between available robots and task partitions are constructed, with the left node being the set of available robots and the right node being the set of task partitions. The element values ​​in the weight matrix correspond to the matching degree between each available robot and each task partition. The KM algorithm is used to solve the maximum weight matching of the bipartite graph based on the matching degree. Each available robot is paired with each task partition in turn to obtain the optimal allocation scheme between available robots and task partitions.

[0012] Preferably, the cloud scheduling platform, based on the optimal allocation scheme between available robots and task zones, packages and transmits all pending goods information for each task zone, along with the path planning from the matching robot to the corresponding task zone, to the cloud scheduling platform. The robot control system in the cloud scheduling platform then issues task instructions to each available robot. After receiving the instructions, the robot autonomously performs the goods picking operation. During the robot's operation, the cloud scheduling platform monitors the robot's location, remaining battery power, goods picking progress, and equipment status in real time, and dynamically updates the robot's scheduling scheme in real time.

[0013] Secondly, the present invention provides a cloud computing-based warehousing and logistics scheduling system for implementing the above-mentioned cloud computing-based warehousing and logistics scheduling method. The system includes a processor, a memory, a communication interface, a cloud scheduling platform, and a warehousing management system. The processor stores computer program instructions for implementing the above-mentioned cloud computing-based warehousing and logistics scheduling method. The cloud scheduling platform includes a robot control system. The communication interface is communicatively connected to the cloud scheduling platform and the warehousing management system.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention utilizes a cloud computing platform to globally optimize task partitioning, path planning, and task allocation, significantly reducing the cost and energy consumption of ineffective robot movement and improving the sorting efficiency and collaborative operation capabilities of robot clusters in warehousing and logistics. Through load and spatial collaborative clustering, it comprehensively evaluates and quantifies the advantages and disadvantages of partitioning, dynamically determining task partitions suitable for the robot scale, balancing the spatial aggregation of goods and load distribution from the source, and avoiding overload or idle robots in a single partition. Based on a gridded map, it generates collision-free paths and accurately calculates various cost characteristics, relying on the computing power of the cloud computing platform to support large-scale parallel path computation, providing realistic and executable cost data for global matching. Through multi-factor global matching, it integrates path, load, and power consumption to construct an objective function, combining the KM algorithm to achieve globally optimal allocation of robots and tasks, overcoming the local limitations of greedy algorithms and ensuring the highest overall system matching degree. Through dynamic scheduling, it monitors the operational status in real time and responds to new orders, faults, and other anomalies, ensuring that the scheduling scheme dynamically adapts to the warehousing environment and guarantees long-term stable system operation. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the implementation of the cloud computing-based warehousing and logistics scheduling method of the present invention. Figure 2 This is a structural block diagram of the cloud computing-based warehousing and logistics scheduling system of the present invention. Detailed Implementation

[0016] The technical features of the present invention will be further described in detail below with reference to the accompanying drawings, so that those skilled in the art can understand them.

[0017] A cloud-based warehouse logistics scheduling method, the implementation process of which is as follows: Figure 1 As shown, the specific implementation steps are as follows: Step S1: Collect information on goods awaiting processing and the status of available robots in the warehouse management system in real time based on the cloud scheduling platform.

[0018] Specifically, a data collection cycle is set. The cloud-based scheduling platform periodically collects goods order information and robot cluster status information from the warehouse management system through an IoT interface in real time. The goods order information includes the quantity of goods to be processed, the three-dimensional spatial coordinates and weight data of each goods to be processed. The robot cluster status information includes the current position coordinates, remaining power, and current load capacity of each robot. The idle robot information is extracted from the robot cluster status information, and the data of robots with remaining power not exceeding the threshold is removed to obtain the number of available robots, as well as the two-dimensional position coordinates, remaining power, and current load capacity data of each available robot. The power threshold can be set to 20% of the rated full power to ensure that the robots participating in the scheduling have the ability to fully execute tasks. The data collection cycle can be set to the second level, such as 5 seconds / time, to dynamically capture changes in robot status, avoid deviations from the actual number of available robots, and avoid data redundancy caused by excessively short cycles, ensuring the efficiency of subsequent clustering iterations. In one embodiment, the warehouse management system receives a batch of newly arrived goods orders. These orders contain N items that need to be picked, i.e., N items to be processed. Each item has its unique three-dimensional spatial coordinates and weight corresponding to its warehouse shelf location. The three-dimensional spatial coordinates of the nth item to be processed are... The weight data is Extract the set of available robots in the warehouse management system that are currently idle and have a remaining battery level of more than 20%. K represents the total number of available robots, and the ID number of the kth available robot is... Its two-dimensional position coordinate data is ( The remaining battery power data is as follows: Current load capacity data is .

[0019] Step S2: Based on the information of the goods to be processed, cluster the goods to be processed and calculate and evaluate the clustering results to obtain the optimal task partitioning scheme.

[0020] Specifically, based on the number of available robots, the range of values ​​for the number of task partitions is set. The number of task partitions cannot exceed the number of available robots. When the total number of available robots is K, the range of values ​​for the number of task partitions is [value missing]. The K-Means++ algorithm is used to select the number of task partitions within a set range as the number of clusters. The coordinate data of the goods to be processed are clustered, and the number of clusters is the same as the number of task partitions selected. Each cluster corresponds to one task partition. Based on the number of task partitions selected and the corresponding clustering results, the mean of the coordinates of all goods to be processed in each task partition is used as the center coordinate of the corresponding task partition. The square Euclidean distance between each goods to be processed in the task partition and its center coordinate is calculated. Based on the number of goods to be processed in the task partition, the square Euclidean distances corresponding to all goods to be processed in the task partition are accumulated as the intra-cluster spatial feature of the corresponding task partition. Based on the number of task partitions selected, the intra-cluster spatial features of all task partitions are accumulated as the intra-cluster spatial compactness corresponding to the value of the number of task partitions. Similarly, the intra-cluster spatial compactness corresponding to each value within the set range of the number of task partitions is obtained. The intra-cluster spatial compactness function is defined as follows: This is used to represent the geographical clustering degree of task partitions. The specific formula for calculating the spatial compactness within a cluster is as follows: In the formula, This indicates that when the number of task partitions is The corresponding intra-cluster space compactness, The value representing the number of task partitions selected. This represents the spatial coordinates of the nth item within the i-th task partition. This represents the center coordinates within the i-th partition, which is the average of the coordinates of all goods to be processed within the i-th task partition. This represents the quantity of goods to be processed within the i-th partition; The squared Euclidean distance between the nth item in the i-th task partition and its center coordinates is used to quantify the spatial deviation between the item to be processed and the center of the task partition. The essence of the spatial compactness within the cluster is the sum of squared errors of the clustering. The smaller the value, the more concentrated the geographical distribution of the items to be processed within the task partition, and the better the spatial compactness.

[0021] Furthermore, based on the selected number of task partitions and the corresponding clustering results, the weight data of all goods to be processed in each task partition is extracted, the total weight of all goods to be processed in each task partition is calculated, the mean of the total weight of all task partitions is taken as the average load, and the square of the difference between the total weight of each task partition and the average load is taken as the inter-cluster load feature of the corresponding task partition. Based on the selected number of task partitions, the inter-cluster load features of all task partitions are accumulated and the ratio between the accumulated value and the number of task partitions is calculated. The square root of the ratio is taken as the inter-cluster load balance corresponding to the value of the number of task partitions. Similarly, the inter-cluster load balance corresponding to each value within the set range of the number of task partitions is obtained. The inter-cluster load balancing function is defined as follows: This is used to quantify the workload differences between task partitions. The specific formula for calculating inter-cluster load balancing is as follows: In the formula, This indicates that when the number of task partitions is The corresponding inter-cluster load balancing degree, The value representing the number of task partitions selected. This represents the total weight of all goods to be processed within the i-th task partition. express The average total weight of all task partitions, i.e., the number of task partitions. The average load over time; the essence of inter-cluster load balancing is the standard deviation of the load of each task partition, which is used to quantify the dispersion of load distribution. The smaller the value, the smaller the difference in workload between task partitions, and the better the load balancing.

[0022] Furthermore, the intra-cluster spatial compactness and inter-cluster load balancing of each value within the set range of the number of task partitions are normalized. Weight coefficients for intra-cluster spatial compactness and inter-cluster load balancing are set based on actual application scenarios. For a given number of task partitions, the sum of the products of the normalized intra-cluster spatial compactness and inter-cluster load balancing values ​​and their corresponding weight coefficients is used as the comprehensive evaluation value for that number of task partitions. Similarly, the comprehensive evaluation value for each value within the set range of the number of task partitions is obtained. The number of task partitions corresponding to the minimum value among all comprehensive evaluation values ​​is taken as the optimal number of task partitions for adapting to the robot scale, and the clustering result corresponding to the optimal number of task partitions is taken as the optimal task partitioning scheme. The comprehensive evaluation function for defining task partitioning is: This is used to comprehensively evaluate the partitioning scheme of the current task by combining space compactness and load balancing. The specific formula for calculating the comprehensive evaluation value is as follows: In the formula, This indicates that when the number of task partitions is The corresponding comprehensive evaluation value at that time The value representing the number of task partitions selected. express The normalization function is expressed as follows: , This indicates that when the number of task partitions is The corresponding intra-cluster space compactness, This indicates that when the number of task partitions is The corresponding inter-cluster load balancing degree, The weighting coefficients represent the spatial compactness within a cluster. This represents the weighting coefficient corresponding to the inter-cluster load balancing degree; Weighting coefficient and It can be configured according to the actual application scenario, for example: it can be set , If reducing robot mobility costs is the priority, then greater attention needs to be paid to the compactness of the cluster space, which can increase... If avoiding robot overload is the priority, then more attention needs to be paid to inter-cluster load balancing, which can increase... The comprehensive evaluation value is the core indicator for quantifying the merits of a task partitioning scheme. The smaller the value, the better the scheme performs in terms of space compactness and load balancing. By determining the optimal task partitioning scheme through the comprehensive evaluation value, the subjectivity of manually setting the number of partitions can be avoided, ensuring that the scheme is adapted to the actual orders and robot scale.

[0023] Step S3: Based on the optimal task partitioning scheme and path planning algorithm, generate the shortest collision-free path from each available robot to each task partition and calculate its path length.

[0024] Specifically, based on the actual warehouse layout, a gridded map containing obstacle information is constructed. The size of each grid cell in the map is set according to the scene's accuracy requirements. Areas that the robot can pass through are marked as passable grids and assigned a value of 1, corresponding to warehouse aisles, work areas, and other areas where the robot can move normally. Areas that the robot cannot pass through are marked as obstacle grids and assigned a value of 1. For areas where robots cannot pass, such as shelves, columns, walls, and equipment, this ensures effective obstacle avoidance during path planning. Based on the optimal task partitioning scheme, the average of the planar coordinates of all goods to be processed within a task partition is used as the planar center coordinates of that task partition. The planar center coordinates of each task partition and the current position coordinates of each available robot are obtained. The current position coordinates of each available robot are used as the path start point, and the center coordinates of each task partition are used as the path end point. The path planning algorithm (A-Star, a path finding and graph traversal algorithm) generates the shortest possible path from the starting point to the ending point of each path, which serves as the shortest collision-free path from the corresponding available robot to each task partition. Based on the gridded map, the sequence of passable grids for each path is obtained. Based on the set grid cell size, the product of the number of passable grids for each path and the grid cell size is used as the path length of the corresponding path, thus obtaining the path length from each available robot to each task partition.

[0025] The size of each grid cell can be set to 0.5m×0.5m. The current position coordinates of the available robots collected in step S1 are used as the starting point of the path, that is, the initial position of each available robot is used as the starting point of the path, and the plane center coordinates of each task partition determined in step S2 are used as the ending point of the path. Since the vertical operation is completed by the robot end alone, the vertical coordinates are ignored and only the plane coordinates are retained for path calculation.

[0026] Step S4: Based on the path length, the information of the goods to be processed, and the status information of the available robots, calculate the path adaptability, load adaptability, and power adequacy between each available robot and each task partition.

[0027] Specifically, based on the path length from each available robot to each task partition, for a given task partition, the maximum and minimum values ​​of the path lengths from all available robots to that task partition are extracted. The difference between the maximum and minimum values ​​is taken as the path baseline difference for that task partition. The difference between the path length from any available robot to that task partition and the minimum value is taken as the path difference for the corresponding available robot. The ratio of the path difference to the path baseline difference is calculated. The result of subtracting the ratio from 1 is taken as the path fit between the corresponding available robot and the task partition. Similarly, the path fit between each available robot and each task partition is obtained. Path adaptability reflects the movement cost of available robots to the task partition. The shorter the path, the higher the matching degree. The formula for calculating path adaptability is as follows: In the formula, This represents the path fit between the k-th available robot and the i-th task partition. This represents the path length from the kth available robot to the i-th task partition. This represents the minimum path length among all available robots moving to each task partition. This represents the maximum path length among all available robots moving to each task partition.

[0028] Furthermore, based on the information of goods to be processed and the status information of available robots, the weight data of all goods to be processed and the current load capacity data of each available robot are extracted. Based on the optimal task partitioning scheme, the maximum value of the weight data of goods to be processed in each task partition is extracted as the maximum load corresponding to each task partition. When the maximum load of a task partition is less than or equal to the current load capacity of an available robot, the ratio between the maximum load and the current load capacity is used as the load adaptability between the corresponding available robot and the task partition. When the maximum load of a task partition is greater than the current load capacity of an available robot, the load adaptability between the corresponding available robot and the task partition is set to 0. Similarly, the load adaptability between each available robot and each task partition is obtained. Load adaptability reflects the compatibility between the current load capacity of the available robot and the weight of the goods to be processed in each task partition. The closer the maximum value of the weight of the goods to be processed in the task partition is to the current load capacity of the available robot, and the less it exceeds the current load capacity of the available robot, the higher the matching degree between the available robot and the task partition. The formula for calculating load adaptability is as follows: In the formula, This represents the load adaptability between the k-th available robot and the i-th task partition. This represents the ID number of the kth available robot. This represents the maximum value among all unprocessed cargo weight data within the i-th task partition. This represents the current load capacity data of the kth available robot.

[0029] Furthermore, based on the status information of available robots, the remaining power data of each available robot is extracted to obtain the power consumption coefficient per unit distance and the basic power consumption of the robot's working state. Based on the path length from each available robot to each task partition, the product of the power consumption coefficient and the path length is added to the basic power consumption to obtain the theoretical power consumption of the corresponding path. Similarly, the theoretical power consumption from each available robot to each task partition is obtained. For a certain task partition, when the remaining power data of the available robot is greater than the theoretical power consumption from the available robot to the task partition, the difference between the remaining power and the theoretical power consumption is calculated. The ratio between the difference and the robot's rated full power is used as the power adequacy between the corresponding available robot and the task partition. When the remaining power data of the available robot is less than or equal to the theoretical power consumption from the available robot to the task partition, the power adequacy between the corresponding available robot and the task partition is set to 0. Similarly, the power adequacy between each available robot and each task partition is obtained. Among them, battery adequacy reflects whether the remaining battery power of the available robot can support path movement and operation power consumption. The more sufficient the battery power, the higher the matching degree. The formula for calculating battery adequacy is as follows: In the formula, This indicates the battery sufficiency between the k-th available robot and the i-th task partition. This represents the remaining battery power of the kth available robot. This represents the theoretical power consumption for the kth available robot to move to the i-th task partition. , This represents the path length from the kth available robot to the i-th task partition. This represents the power consumption coefficient per unit distance of the robot. Basic power consumption indicating the robot's operating status. This indicates the robot's rated full charge.

[0030] Step S5: Calculate the matching degree between each available robot and each task partition based on the three characteristics, and determine the optimal allocation scheme between available robots and task partitions.

[0031] Specifically, weight coefficients for path adaptability, load adaptability, and power adequacy are set according to actual application requirements, with the sum of each weight coefficient being 1. Based on the path adaptability, load adaptability, and power adequacy between each available robot and each task partition, the sum of the products of the three characteristics and their respective weight coefficients is used as the matching degree between the corresponding available robot and the task partition. When the load adaptability or power adequacy between an available robot and a task partition is 0, the matching degree between the corresponding available robot and the task partition is set to 0, thus obtaining the matching degree between each available robot and each task partition. A bipartite graph and weight matrix are constructed between available robots and task partitions, with the left node representing the set of available robots and the right node representing the set of task partitions. The element values ​​in the weight matrix correspond to the matching degree between each available robot and each task partition. The KM algorithm (Kuhn-Munkres, the maximum weight matching algorithm for bipartite graphs) is used to solve the maximum weight matching of the bipartite graph based on the matching degree. Each available robot is paired with each task partition in turn to obtain the optimal allocation scheme between available robots and task partitions. The formula for calculating the matching degree is as follows: In the formula, This represents the matching degree between the k-th available robot and the i-th task partition. This represents the path fit between the k-th available robot and the i-th task partition. This represents the load adaptability between the k-th available robot and the i-th task partition. This indicates the battery sufficiency between the k-th available robot and the i-th task partition. , , These are the weighting coefficients for path adaptability, load adaptability, and battery adequacy, respectively, to meet the requirements. + + The weighting coefficients are set according to the business priorities of the actual application scenario. Generally, path cost has a higher priority in warehousing scenarios and can be set accordingly. , , If load adaptability Or sufficient battery power If any item in the list is 0, it indicates that the load is overloaded or the power is insufficient. In this case, let... The combination between the kth available robot and the i-th task partition is directly excluded from the matching range.

[0032] In addition, when the number of available robots exceeds the number of task partitions, unmatched available robots are put into standby or charging mode. Unmatched available robots wait for scheduling instructions as substitutes and do not participate in the scheduling scheme of the current task to avoid confusion in robot running routes.

[0033] Step S6: Issue task instructions based on the optimal allocation scheme, schedule available robots and monitor their operation status in real time.

[0034] Specifically, the cloud-based scheduling platform, based on the optimal allocation scheme between available robots and task zones, packages and transmits all pending goods information for each task zone, along with the path planning from the matching robot to the corresponding task zone, to the cloud-based scheduling platform. The robot control system in the cloud-based scheduling platform then issues task instructions to each available robot. After receiving the instructions, the robot autonomously executes the goods picking operation. During the robot's operation, the cloud-based scheduling platform monitors the robot's location, remaining battery power, goods picking progress, and equipment status in real time, and dynamically updates the robot's scheduling scheme in real time.

[0035] In addition, when abnormal situations such as new orders being inserted, robots completing tasks, robot malfunctions, or path blockages occur, a new round of scheduling processes is immediately triggered to achieve real-time response to dynamic changes in the scenario and ensure overall operational efficiency.

[0036] This invention also discloses a cloud computing-based warehousing and logistics scheduling system, the structure of which is as follows: Figure 2 As shown, the warehousing and logistics scheduling method based on cloud computing described above includes: a processor, a memory, a communication interface, a cloud scheduling platform, and a warehousing management system. The processor stores computer program instructions for implementing the cloud computing-based warehousing and logistics scheduling method described above. The cloud scheduling platform includes a robot control system, and the communication interface is connected to both the cloud scheduling platform and the warehousing management system.

[0037] The embodiments included in this invention are merely descriptions of preferred embodiments of the invention and are not limited to the precise structures described above and shown in the accompanying drawings. Various modifications and changes can be made without departing from the scope of protection. Any variations and improvements made by those skilled in the art to the technical solutions of this invention without departing from the design concept of this invention should fall within the scope of protection of this invention.

Claims

1. A cloud computing-based warehouse logistics scheduling method, characterized in that: The system collects information on goods to be processed and the status of available robots in the warehouse management system in real time through a cloud-based scheduling platform. Based on the information on goods to be processed, the system clusters the goods and calculates and evaluates the clustering results to obtain the optimal task partitioning scheme. Based on the optimal task partitioning scheme and path planning algorithm, the system generates the shortest collision-free path from each available robot to each task partition and calculates its path length. Based on path length, information on goods to be processed, and status information of available robots, the path adaptability, load adaptability, and power adequacy of each available robot with each task partition are calculated. The matching degree between each available robot and each task partition is calculated based on three characteristics to determine the optimal allocation scheme between available robots and task partitions; task instructions are issued based on the optimal allocation scheme to schedule available robots and monitor their operation status in real time. In this process, weight coefficients for path adaptability, load adaptability, and power adequacy are set according to actual application requirements. The sum of each weight coefficient is 1. The sum of the products of the three characteristics and their respective weight coefficients is used as the matching degree between the corresponding available robot and the task partition. When the load adaptability or power adequacy between an available robot and a task partition is 0, the matching degree between the corresponding available robot and the task partition is set to 0.

2. The cloud computing-based warehousing and logistics scheduling method according to claim 1, characterized in that, The method of collecting information on goods awaiting processing and the status information of available robots in the warehouse management system in real time based on the cloud scheduling platform includes: setting a data collection period; the cloud scheduling platform periodically collects goods order information and robot cluster status information from the warehouse management system in real time through the Internet of Things interface; the goods order information includes the quantity of goods awaiting processing, the three-dimensional spatial coordinate data and weight data of each goods awaiting processing; the robot cluster status information includes the current position coordinates, remaining power and current load capacity of each robot; extracting the information data of idle robots from the robot cluster status information and removing the data of robots whose remaining power does not exceed the threshold, to obtain the number of available robots, and the two-dimensional position coordinate data, remaining power data and current load capacity data of each available robot.

3. The cloud computing-based warehousing and logistics scheduling method according to claim 2, characterized in that, The process of clustering the goods to be processed based on their information and evaluating the clustering results includes: setting a range for the number of task partitions based on the number of available robots, ensuring that the number of task partitions does not exceed the number of available robots; using the K-Means++ algorithm, sequentially selecting the number of task partitions within the set range as the number of clusters; clustering the coordinate data of the goods to be processed, resulting in the same number of clusters as the selected number of task partitions, with each cluster corresponding to one task partition; based on the selected number of task partitions and the corresponding clustering results, using the mean of all coordinates of the goods to be processed within each task partition as the center coordinate of the corresponding task partition; calculating the squared Euclidean distance between each goods to be processed within the task partition and its center coordinate; summing the squared Euclidean distances of all goods to be processed within the task partition based on the number of goods to be processed within the task partition as the intra-cluster spatial feature of the corresponding task partition; summing the intra-cluster spatial features of all task partitions based on the selected number of task partitions as the intra-cluster spatial compactness corresponding to the set value of the number of task partitions; similarly, obtaining the intra-cluster spatial compactness corresponding to each value within the set range of the number of task partitions.

4. The cloud computing-based warehousing and logistics scheduling method according to claim 3, characterized in that, The calculation and evaluation of the clustering results includes: based on the selected number of task partitions and the corresponding clustering results, extracting the weight data of all goods to be processed in each task partition, calculating the total weight of all goods to be processed in each task partition, taking the mean of the total weight of all task partitions as the average load, taking the square of the difference between the total weight and the average load of each task partition as the inter-cluster load feature of the corresponding task partition, based on the selected number of task partitions, accumulating the inter-cluster load features of all task partitions and calculating the ratio between the accumulated value and the number of task partitions, taking the square root of the ratio as the inter-cluster load balance corresponding to the number of task partitions, and similarly obtaining the inter-cluster load balance corresponding to each value within the set range of the number of task partitions.

5. The cloud computing-based warehousing and logistics scheduling method according to claim 4, characterized in that, The calculation and evaluation of the clustering results also includes: normalizing the intra-cluster spatial compactness and inter-cluster load balance corresponding to each value within the set range of the number of task partitions; setting weight coefficients for intra-cluster spatial compactness and inter-cluster load balance based on the actual application scenario; for a certain value of the number of task partitions, summing the products of the normalized values ​​of the intra-cluster spatial compactness and inter-cluster load balance with the corresponding weight coefficients, and using this sum as the comprehensive evaluation value corresponding to that value of the number of task partitions; similarly, obtaining the comprehensive evaluation value corresponding to each value within the set range of the number of task partitions; taking the number of task partitions corresponding to the minimum value among all comprehensive evaluation values ​​as the optimal number of task partitions adapted to the robot scale; and taking the clustering result corresponding to the optimal number of task partitions as the optimal task partitioning scheme.

6. The cloud computing-based warehousing and logistics scheduling method according to claim 2, characterized in that, The method based on the optimal task partitioning scheme and path planning algorithm generates the shortest collision-free path from each available robot to each task partition and calculates its path length. This includes: constructing a gridded map containing obstacle information based on the actual warehouse layout; setting the size of each grid cell in the map according to the scene's accuracy requirements; marking areas that the robot can normally pass through as passable grids and assigning a value of 1; and marking areas that the robot cannot pass through as obstacle grids and assigning a value of 1. Based on the optimal task partitioning scheme, the average of the planar coordinates of all unprocessed goods within a task partition is used as the planar center coordinates of the corresponding task partition. The planar center coordinates of each task partition and the current position coordinates of each available robot are obtained. The current position coordinates of each available robot are used as the path start point, and the center coordinates of each task partition are used as the path end point. The path planning algorithm generates the shortest path from the starting point to the end point of each path, which serves as the shortest collision-free path from the corresponding available robot to each task partition. Based on the gridded map, the sequence of passable grids for each path is obtained. Based on the set grid cell size, the product of the number of passable grids for each path and the grid cell size is used as the path length of the corresponding path, thus obtaining the path length from each available robot to each task partition.

7. The cloud computing-based warehousing and logistics scheduling method according to claim 2, characterized in that, The calculation of path adaptability, load adaptability, and power adequacy between each available robot and each task partition based on path length, information on goods to be processed, and status information of available robots includes: based on the path length from each available robot to each task partition, for a certain task partition, extracting the maximum and minimum values ​​of the path lengths from all available robots to that task partition, using the difference between the maximum and minimum values ​​as the path baseline difference for that task partition, using the difference between the path length from any available robot to that task partition and the minimum value as the path difference for the corresponding available robot, calculating the ratio of the path difference to the path baseline difference, and subtracting the ratio from 1 as the path adaptability between the corresponding available robot and that task partition. Similarly, the path adaptability between each available robot and each task partition is obtained. Based on the information of goods to be processed and the status information of available robots, the weight data of all goods to be processed and the current load capacity data of each available robot are extracted. Based on the optimal task partitioning scheme, the maximum value of the weight data of goods to be processed in each task partition is extracted as the maximum load corresponding to each task partition. When the maximum load of a task partition is less than or equal to the current load capacity of an available robot, the ratio between the maximum load and the current load capacity is used as the load adaptability between the corresponding available robot and the task partition. When the maximum load of a task partition is greater than the current load capacity of an available robot, the load adaptability between the corresponding available robot and the task partition is set to 0. Similarly, the load adaptability between each available robot and each task partition is obtained. Based on the status information of available robots, the remaining power data of each available robot is extracted to obtain the power consumption coefficient per unit distance and the basic power consumption of the robot in operation. Based on the path length from each available robot to each task partition, the product of the power consumption coefficient and the path length is added to the basic power consumption to obtain the theoretical power consumption of the corresponding path. Similarly, the theoretical power consumption from each available robot to each task partition is obtained. For a certain task partition, when the remaining power data of the available robot is greater than the theoretical power consumption from the available robot to the task partition, the difference between the remaining power and the theoretical power consumption is calculated. The ratio between the difference and the robot's rated full power is used as the power adequacy between the corresponding available robot and the task partition. When the remaining power data of the available robot is less than or equal to the theoretical power consumption from the available robot to the task partition, the power adequacy between the corresponding available robot and the task partition is set to 0. Similarly, the power adequacy between each available robot and each task partition is obtained.

8. The cloud computing-based warehousing and logistics scheduling method according to claim 1, characterized in that, The process of calculating the matching degree between each available robot and each task partition based on three characteristics, and determining the optimal allocation scheme between available robots and task partitions, includes: obtaining the matching degree between each available robot and each task partition based on path adaptability, load adaptability, and power sufficiency; constructing a bipartite graph and weight matrix between available robots and task partitions, where the left node is the set of available robots and the right node is the set of task partitions; the element values ​​in the weight matrix correspond to the matching degree between each available robot and each task partition; using the KM algorithm to solve for the maximum weight matching of the bipartite graph based on the matching degree; and sequentially pairing each available robot with each task partition to obtain the optimal allocation scheme between available robots and task partitions.

9. The cloud computing-based warehousing and logistics scheduling method according to claim 1, characterized in that, The process of issuing task instructions based on the optimal allocation scheme, scheduling available robots, and monitoring their operational status in real time includes: the cloud scheduling platform, based on the optimal allocation scheme between available robots and task partitions, uniformly packages and transmits all pending goods information for each task partition, along with the path planning from the matching robot to the corresponding task partition, to the cloud scheduling platform; the robot control system in the cloud scheduling platform issues task instructions to each available robot; and the robot autonomously executes goods picking operations after receiving the instructions. During the robot's operation, the cloud scheduling platform monitors the robot's location, remaining battery power, goods picking progress, and equipment status in real time, and dynamically updates the robot's scheduling scheme in real time.

10. A cloud computing-based warehousing and logistics scheduling system, characterized in that: The system includes a processor, a memory, a communication interface, a cloud scheduling platform, and a warehouse management system. The processor stores computer program instructions for implementing the cloud-based warehouse logistics scheduling method according to any one of claims 1 to 9. The cloud scheduling platform includes a robot control system. The communication interface is communicatively connected to the cloud scheduling platform and the warehouse management system.

Images