Dynamic resource priority intelligent allocation method based on device dispatching system

Abstract

This invention relates to the field of equipment scheduling technology and discloses a dynamic resource priority intelligent allocation method based on an equipment scheduling system, comprising the following steps: S101, collecting equipment status data and resource status data; S102, classifying tasks and calculating task urgency; S103, calculating a resource heat index by combining the idleness of each on-site execution resource with the moving average occupancy rate within a sliding time window; S104, calculating a spatiotemporal potential value based on the shortest feasible path distance between the task and the on-site execution resource, the resource heat index, and the current cumulative waiting time; S105, constructing a fit matrix based on the task urgency, resource heat index, and spatiotemporal potential value; S106, performing optimization calculations to obtain the final allocation result of the task and the on-site execution resource. This invention achieves intelligent matching of tasks and resources driven by multiple factors, significantly improving scheduling efficiency and system response capability.

Description

Technical Field

[0001] This invention belongs to the field of equipment scheduling technology, specifically relating to a dynamic resource priority intelligent allocation method based on an equipment scheduling system. Background Technology

[0002] In modern production, warehousing, and logistics, efficient resource allocation plays a crucial role in improving overall operational efficiency and reducing costs. However, with the continuous expansion of the industry and the increasing complexity of business operations, traditional resource allocation methods are gradually revealing numerous drawbacks.

[0003] On the one hand, in the task allocation process, the past reliance on manual experience or simple rule-based algorithms makes it difficult to accurately assess the urgency of tasks, the real-time status of resources, and the compatibility between the two. For example, in warehouse picking tasks, facing a large number of orders of different types and with different deadlines, traditional methods cannot quickly and reasonably allocate tasks to the most suitable picking personnel or equipment, leading to frequent problems such as task backlog and delayed delivery. On the other hand, resource management lacks dynamism and refinement. The actual load capacity and idle status of on-site execution resources such as equipment and manpower are constantly changing, and traditional resource scheduling is difficult to track and adjust flexibly in real time. Taking stacker cranes in automated warehouses as an example, when the demand for goods entering and leaving a certain area suddenly increases, traditional scheduling cannot detect and reasonably allocate stacker crane resources in time, easily resulting in some stacker cranes being overworked while others are idle, leading to low resource utilization. In addition, existing scheduling systems often fail to take into account multiple factors; in pursuing task execution efficiency, resource consumption may be ignored; in focusing on resource utilization, urgent task needs may not be guaranteed. Summary of the Invention

[0004] This invention provides a dynamic resource priority intelligent allocation method based on an equipment scheduling system, which solves the technical problem in related technologies that cannot simultaneously take into account task timeliness, resource load status and spatial reachability, resulting in inaccurate scheduling matching, low response efficiency and poor resource utilization.

[0005] This invention provides a dynamic resource priority intelligent allocation method based on an equipment scheduling system, comprising the following steps:

[0006] S101, collect equipment status data and resource status data. The equipment status data includes: equipment location, idle time, failure probability, and task queue length. The resource status data includes: task information, storage location information, and personnel information. The task information includes: task type and task deadline.

[0007] On-site execution resources consist of equipment status data, cargo location information, and personnel information, with each on-site execution resource corresponding to a unique location coordinate.

[0008] S102, classify tasks based on task type and task deadline, compare the remaining time of each task with the historical average remaining time of the same task type, and calculate the task urgency by combining the preset penalty coefficient.

[0009] S103 divides the field execution resources into equipment layer, storage layer and human-machine collaboration layer, and further subdivides them according to physical lanes. Combining the idleness of each field execution resource with the moving average occupancy rate within the sliding time window, the resource heat index is calculated.

[0010] S104, based on the voxel mesh structure preset in the three-dimensional space of the warehouse, marks the voxel units in it, and calculates the spatiotemporal potential energy value based on the shortest feasible path distance between the voxel where the target location of the task is located and the voxel where the on-site execution resource is located, the resource heat index and the current cumulative waiting time of the task.

[0011] S105, construct the compatibility matrix between the task and the on-site execution resources based on the task urgency, resource heat index and spatiotemporal potential energy value;

[0012] S106, perform optimization calculations to maximize the total fitness score of the fitness matrix, and obtain the final allocation result of the task and on-site execution resources.

[0013] Furthermore, the calculation process for task urgency includes:

[0014] S201, extract the task type, task deadline, and current time, establish a task type index, and determine the time base of the task;

[0015] S202, take the difference between the task deadline and the current time as the remaining time of the task, and obtain the historical average remaining time under the same task type;

[0016] S203, calculate the ratio of the remaining time of the task to the historical average remaining time to obtain the time pressure comparison coefficient;

[0017] S204. Multiply the preset penalty coefficient by the time pressure comparison coefficient to obtain the task urgency.

[0018] Furthermore, the on-site execution resources are divided into equipment layer, storage layer and human-machine collaboration layer according to function, and the resource units are further subdivided within each layer according to physical roadway number;

[0019] Within a set sliding window, the occupancy status of resource units is continuously recorded to obtain the occupancy duration. The ratio of the occupancy duration to the duration of the sliding window is calculated to obtain the moving average occupancy rate of the resource unit.

[0020] The resource heat index is calculated by weighting and integrating the difference between the availability and the occupancy rate with the moving average occupancy rate according to preset weights.

[0021] Furthermore, based on the pre-defined voxel mesh structure in the warehouse's three-dimensional space, the voxel units within it are labeled, including:

[0022] S301, the warehouse three-dimensional space is rasterized according to the preset voxel size, and a continuously numbered voxel index is generated;

[0023] S302, Read the location coordinates of the on-site execution resource and map the on-site execution resource to the corresponding voxel index;

[0024] S303, mark each voxel unit with a pass tag. The pass tags include: fixed obstacle tag, temporary occupation tag and passable tag.

[0025] Furthermore, the process of obtaining the spatiotemporal potential energy value includes:

[0026] S401, based on the voxel where the task is located and the voxel where the on-site execution resources are located, uses the Dijkstra algorithm to obtain the shortest feasible path distance;

[0027] S402, by calculating the difference between the current time and the task creation time, the current cumulative waiting time of the task is obtained;

[0028] S403, calculate the average distance of the shortest feasible path in the past thirty days as the distance baseline value, and calculate the average current cumulative waiting time of the task in the past thirty days as the time baseline value;

[0029] S404, multiply the ratio of the shortest feasible path distance to the distance benchmark value by the first preset weight to obtain the distance term, multiply the resource heat index by the second preset weight to obtain the resource heat term, multiply the ratio of the current cumulative waiting time to the time benchmark value by the third preset weight to obtain the time term, and sum the distance term and the resource heat term and then subtract the time term to obtain the spatiotemporal potential energy value. The values ​​of the first preset weight, the second preset weight and the third preset weight are in the range of 0 to 1, and their sum is 1.

[0030] Furthermore, a suitability matrix between the task and on-site execution resources is constructed based on task urgency, resource heat index, and spatiotemporal potential energy value, including:

[0031] S501 linearly standardizes the task urgency and resource heat index to obtain standardized urgency and standardized heat.

[0032] S502, calculate the ratio of the sum of the standardized urgency and the standardized popularity of OnePlus to obtain the benchmark matching value;

[0033] S503: For each pair of tasks and on-site execution resources, the spatiotemporal potential energy value is taken as the inverse number and used as the exponent to obtain the exponential decay term. The exponential decay term is then multiplied by the benchmark matching value to obtain the fit score.

[0034] S504. Write the fit score into the corresponding position of the task and the on-site execution resources to complete the fit matrix filling.

[0035] Furthermore, optimization operations are performed to maximize the total fitness score of the fitness matrix, resulting in the final allocation of tasks and on-site execution resources, including:

[0036] S601, Read the fitness matrix, sum the fitness scores of all matrix elements, and set it as the objective function;

[0037] S602 sets task uniqueness constraints and resource capacity constraints to limit the space of feasible matching solutions;

[0038] S603 performs optimization operations under the objective function and constraints to obtain the final allocation result.

[0039] Furthermore, the optimization operation includes:

[0040] S701, traverse from high to low according to the suitability score, select on-site execution resources with available capacity for unassigned tasks, and generate an initial matching table;

[0041] S702, perform on-site resource exchange within the initial matching table. If the total matching score increases after the exchange, retain the exchange and continue iterating until the total matching score no longer increases after two consecutive iterations.

[0042] S703 outputs the final matching table as the final allocation result. The final matching table records the unique on-site execution resource corresponding to each task.

[0043] Furthermore, it also includes the following instruction issuance steps:

[0044] S801, read the final allocation result, generate a scheduling instruction package for each pair of tasks and on-site execution resources. The instruction package fields include task identifier, resource identifier, starting position coordinates and target position coordinates;

[0045] S802, the scheduling instruction packet is sent to the corresponding field execution resource, and the field execution resource is waited for the return of the receipt confirmation information;

[0046] S803 records the command status as sent after receiving the confirmation message. If no confirmation is received within the preset timeout period, it is marked as pending retry.

[0047] The beneficial effects of this invention are as follows: By introducing three indicators—task urgency, resource heat index, and spatiotemporal potential energy value—this invention comprehensively considers task timeliness pressure, resource load status, and spatial reachability, achieving final matching of tasks and resources driven by multiple factors. It dynamically quantifies task urgency by comparing task type indexes with historical data; it accurately reflects resource busyness by integrating real-time occupancy with historical trends; and it precisely calculates the feasible distance and traversal difficulty between tasks and resources by combining voxel grid structures and path planning algorithms. Finally, by constructing a fitness matrix and performing constrained optimization operations, it outputs allocation results that meet business requirements. Compared to traditional static scheduling strategies, this invention can significantly improve resource utilization, task response efficiency, and the overall operational stability of the scheduling system. Attached Figure Description

[0048] Figure 1 This is a flowchart of the dynamic resource priority intelligent allocation method based on the equipment scheduling system of the present invention. Detailed Implementation

[0049] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0050] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0051] like Figure 1 As shown, the dynamic resource priority intelligent allocation method based on the equipment scheduling system includes the following steps:

[0052] S101, collect equipment status data and resource status data. The equipment status data includes: equipment location, idle time, failure probability, and task queue length. The resource status data includes: task information, storage location information, and personnel information. The task information includes: task type and task deadline.

[0053] On-site execution resources consist of equipment status data, cargo location information, and personnel information, with each on-site execution resource corresponding to a unique location coordinate.

[0054] S102, classify tasks based on task type and task deadline, compare the remaining time of each task with the historical average remaining time of the same task type, and calculate the task urgency by combining the preset penalty coefficient.

[0055] S103 divides the field execution resources into equipment layer, storage layer and human-machine collaboration layer, and further subdivides them according to physical lanes. Combining the idleness of each field execution resource with the moving average occupancy rate within the sliding time window, the resource heat index is calculated.

[0056] S104, based on the voxel mesh structure preset in the three-dimensional space of the warehouse, marks the voxel units in it, and calculates the spatiotemporal potential energy value based on the shortest feasible path distance between the voxel where the target location of the task is located and the voxel where the on-site execution resource is located, the resource heat index and the current cumulative waiting time of the task.

[0057] S105, construct the compatibility matrix between the task and the on-site execution resources based on the task urgency, resource heat index and spatiotemporal potential energy value;

[0058] S106, perform optimization calculations to maximize the total fitness score of the fitness matrix, and obtain the final allocation result of the task and on-site execution resources.

[0059] In one embodiment of the present invention, the equipment status data is obtained in real time by connecting to the warehouse equipment control system. The idle degree is determined by detecting whether the equipment is currently executing a task. If the equipment is not executing any task, it is marked as idle. The idle degree is calculated by combining the idle time ratio per unit time. The failure probability is determined by a comprehensive assessment of the equipment's historical failure records, real-time operating parameters, and failure warning thresholds provided by the manufacturer. The task queue length is obtained by statistically analyzing the number of tasks that the equipment has received but not executed, as recorded by the equipment control system.

[0060] For resource status data, task information is obtained by connecting to the warehouse management system. Task types include: order picking, inventory replenishment, and product shelving. The task deadline is preset by the system according to order requirements or business rules. Location information is extracted from the location management module of the warehouse management system, including location number, location area, and corresponding stored product information. Personnel information is collected through personnel handheld terminals or workstation check-in systems, recording the current location, work group, and work status of the personnel involved in the operation.

[0061] In one embodiment of the present invention, the calculation of task urgency is achieved through comparative analysis of task time characteristics and historical data, specifically including:

[0062] S201, extract the task type, task deadline, and current time, establish a task type index, and determine the time base of the task;

[0063] S202, take the difference between the task deadline and the current time as the remaining time of the task, and obtain the historical average remaining time under the same task type; specifically, retrieve the historical task remaining time data of the same task type from the historical task database, and obtain the historical average remaining time of the task type through statistical calculation, which serves as a reference benchmark for measuring the time pressure of the current task.

[0064] S203, calculate the ratio of the remaining time of the task to the historical average remaining time to obtain the time pressure comparison coefficient; if the coefficient is less than 1, it indicates that the remaining time of the current task is lower than the historical average of the same type, and the time pressure is relatively high; if it is greater than 1, it indicates that the remaining time of the current task is relatively sufficient, and the time pressure is relatively low.

[0065] S204, multiply the preset penalty coefficient by the time pressure comparison coefficient to obtain the task urgency; the formula for calculating the task urgency is: , Indicates task urgency, reflecting the difference between the actual remaining time for the task and the historical performance of similar tasks. Indicates the task deadline. Indicates the current time. Indicates the historical average remaining time. This represents the preset penalty coefficient, used to strengthen the weight of the impact of event pressure on urgency.

[0066] In one embodiment of the present invention, the field execution resources are divided into a device layer, a storage layer and a human-machine collaboration layer according to their functions, and the resource units are further subdivided within each layer according to the physical tunnel number.

[0067] The equipment layer includes various execution devices such as AGVs and robotic arms; the storage layer includes storage facilities such as shelves and containers; and the human-machine collaboration layer includes areas requiring human-machine cooperation, such as picking workstations and verification stations. Within each layer, resource units are further subdivided according to physical lane numbers to make the spatial positioning of resource units more accurate, providing a foundation for subsequent location-based scheduling optimization.

[0068] Within a set sliding window, the occupancy status of resource units is continuously recorded to obtain the occupancy duration. The ratio of the occupancy duration to the duration of the sliding window is calculated to obtain the moving average occupancy rate of the resource unit.

[0069] The sliding window is set to 5 minutes. Within this window, the occupancy status of each resource unit is continuously recorded, such as whether the equipment is performing a task or whether the storage space is occupied. The occupancy time of the resource unit within the sliding window is accumulated, and then the moving average occupancy rate of the resource unit is obtained. This occupancy rate can smooth out short-term fluctuations and more accurately reflect the actual load level of the resource unit.

[0070] The resource heat index is calculated by weighting and integrating the value minus the idle degree and the moving average occupancy rate according to preset weights. The value minus the idle degree represents the current occupancy of the resource unit, while the moving average occupancy rate reflects the historical load trend. The higher the value of the weighted resource heat index, the busier the resource unit is and the lower the schedulable redundancy.

[0071] Through the above process, the structured layering and spatial subdivision of resources are realized, which improves the management accuracy. Furthermore, by integrating the moving average occupancy rate with the current occupancy level, the load status of resources is reflected, so that the resource heat index can reflect both the real-time busy level and the historical load trend, providing a quantitative basis for the accurate matching of subsequent tasks and resources.

[0072] In one embodiment of the present invention, based on a pre-defined voxel mesh structure in the three-dimensional space of the warehouse, the voxel units therein are marked, aiming to transform the physical space into a quantifiable and calculable digital space, providing a spatial basis for subsequent path planning and resource scheduling. The specific implementation process includes:

[0073] S301, the three-dimensional space of the warehouse is rasterized according to the preset voxel size, and a continuously numbered voxel index is generated; wherein, the preset voxel size can be set to 0.5 cubic meters / voxel; by dividing the length, width and height of the warehouse equally according to the voxel size, the entire warehouse space is discretized into several regular cubic voxel units, and each voxel unit is assigned a unique continuous number as a voxel index, thereby realizing the digital mapping of the physical space of the warehouse, and transforming the originally abstract spatial location into index information that can be directly called.

[0074] S302, read the location coordinates of the field execution resource and map the field execution resource to the voxel index corresponding to its voxel unit; this process establishes the association between the field execution resource and the digital spatial unit, so that the location information of the resource can be quickly identified and called in the form of voxel index, providing a unified reference standard for subsequent spatial location-based calculations;

[0075] In step S303, each voxel unit is labeled with access tags, including fixed obstacle tags, temporary occupation tags, and passable tags. Fixed obstacle tags mark voxel units containing immovable fixed facilities within the warehouse, indicating that the area cannot be used as a passageway. Temporary occupation tags mark voxel units that are temporarily inaccessible due to temporary operations; their occupation status changes dynamically as the operations progress. Passable tags mark voxel units that are currently unobstructed and can be used as normal passageways. These tags clearly define the accessibility attributes of each voxel unit, providing accurate spatial constraint information for subsequent path planning algorithms.

[0076] Through the above steps, this invention realizes the digital and structured representation of the three-dimensional space of the warehouse. On the one hand, the discretization of the voxel grid reduces the complexity of spatial calculation, making subsequent operations such as path distance calculation and resource location association easier and more efficient to execute. On the other hand, the labeling of voxel units provides the system with clear spatial access rules, ensuring the rationality and feasibility of path planning, and thus providing accurate spatial basic data for the spatiotemporal matching of tasks and resources, improving the spatial adaptability and overall efficiency of resource scheduling.

[0077] In one embodiment of the present invention, the process of obtaining the spatiotemporal potential energy value includes:

[0078] S401, based on the voxel where the task is located and the voxel where the on-site execution resources are located, uses the Dijkstra algorithm to obtain the shortest feasible path distance; this path distance uses the voxel grid as the spatial reference and combines the access labels of the voxel units to ensure that the path planning conforms to the actual access constraints of the warehouse. The obtained shortest feasible path distance intuitively reflects the degree of physical spatial obstruction between the task and the resources.

[0079] S402, by calculating the difference between the current time and the task creation time, the current cumulative waiting time of the task is obtained;

[0080] S403, calculate the average distance of the shortest feasible path in the past thirty days as the distance baseline value, and calculate the average current cumulative waiting time of the task in the past thirty days as the time baseline value;

[0081] S404, multiply the ratio of the shortest feasible path distance to the distance benchmark value by the first preset weight to obtain the distance term, multiply the resource heat index by the second preset weight to obtain the resource heat term, multiply the ratio of the current cumulative waiting time to the time benchmark value by the third preset weight to obtain the time term, and sum the distance term and the resource heat term and then subtract the time term to obtain the spatiotemporal potential energy value. The values ​​of the first preset weight, the second preset weight and the third preset weight are in the range of 0 to 1, and their sum is 1.

[0082] The formula for calculating the spatiotemporal potential energy is as follows: , This represents the spatiotemporal potential energy value, used to quantify the spatiotemporal matching cost between the task and on-site execution resources. , and These represent the first preset weight, the second preset weight, and the third preset weight, respectively. This represents the distance of the shortest feasible path. Indicates the distance to the reference value. Indicates the resource heat index, This indicates the current cumulative waiting time. Indicates the time base value.

[0083] The spatiotemporal potential energy value calculated through the above process not only avoids the one-sidedness of judging the rationality of space solely by distance, but also correlates the actual load capacity of resources through the resource heat index, and reflects the time urgency of the task by combining the waiting time. This provides a comprehensive spatiotemporal cost basis for the subsequent task and resource suitability calculation, which helps to promote the tilt of tasks towards resources that are close, have low load, and have long waiting time, thereby improving the overall spatial efficiency and time response speed of scheduling.

[0084] In one embodiment of the present invention, a suitability matrix between the task and on-site execution resources is constructed based on task urgency, resource heat index, and spatiotemporal potential energy value, including:

[0085] S501 linearly standardizes the task urgency and resource heat index to obtain standardized urgency and standardized heat. Specifically, the max-min normalization method can be used to map the task urgency and resource heat index to the interval between 0 and 1, eliminating the influence of the difference in the absolute value of the urgency of different types of tasks, and making the resource heat index of different resources comparable.

[0086] S502, calculate the ratio of the sum of the standardized urgency and the standardized popularity of OnePlus to obtain the benchmark matching value;

[0087] S503: For each pair of tasks and on-site execution resources, the spatiotemporal potential energy value is taken as the inverse number and used as the exponent to obtain the exponential decay term. The exponential decay term is then multiplied by the benchmark matching value to obtain the fit score.

[0088] The formula for calculating the fit score is as follows: , This represents the fit score between the i-th task and the j-th on-site execution resource. The higher the score, the better the match between the two. This represents the standardized urgency of the i-th task. This represents the standardization popularity of the j-th on-site execution resource. This represents the spatiotemporal potential energy value between the i-th task and the j-th on-site execution resource, where i represents the task index and j represents the on-site execution resource index.

[0089] S504. Write the fit score into the position corresponding to the task and the on-site execution resource to complete the fit matrix filling; where the rows of the matrix represent tasks and the columns represent on-site execution resources, and the value of each matrix element reflects the degree of matching between the corresponding task and the on-site execution resource.

[0090] In one embodiment of the present invention, an optimization operation is performed to maximize the total fitness score of the fitness matrix, resulting in the final allocation of tasks and on-site execution resources, including:

[0091] S601, Read the fitness matrix, sum the fitness scores of all matrix elements, and set it as the objective function; Specifically, by summing the fitness scores of all elements in the matrix, the total fitness score representing the overall matching effect is obtained, and the total fitness score is set as the objective function of the optimization operation. That is, the core objective of the optimization is to maximize the total fitness score, thereby achieving the overall optimal matching between the task and the on-site execution resources.

[0092] S602 sets task uniqueness constraints and resource capacity constraints to limit the feasible matching solution space. It should be noted that the above constraints are set to ensure the feasibility and rationality of the allocation results. Among them, the task uniqueness constraint stipulates that each task can only be assigned to a unique field execution resource to avoid duplicate assignment or omission of tasks. The resource capacity constraint limits the total number of tasks that each field execution resource can handle to its maximum capacity based on the actual carrying capacity of the field execution resources, to prevent the field execution resources from affecting the execution efficiency due to overload.

[0093] S603, Perform optimization operations under the objective function and constraints to obtain the final allocation result; wherein, the optimization operations include:

[0094] S701: Iterate through all matching combinations of tasks and on-site execution resources from high to low according to the suitability score, and select on-site execution resources with remaining capacity for unassigned tasks in turn to generate an initial matching table; that is, when a task has a high suitability score with an on-site execution resource, the task is assigned to that resource first, until all tasks are assigned or the resources reach their capacity limit; this process quickly builds a basic allocation scheme through a greedy strategy to ensure that high-suitability task-resource combinations are reserved first.

[0095] S702: Within the current matching table, select two on-site execution resources and attempt to swap some of their assigned tasks. Calculate the change in the total fit score after the swap. If the total fit score after the swap is higher than before, retain the swap result and update the matching table. Repeat the above swapping and evaluation process until the total fit score does not improve in two consecutive iterations, indicating that the current matching table has reached a local optimum, and stop the iteration. This process gradually optimizes the allocation scheme through local adjustments, compensating for the potential local optima limitations of the initial greedy strategy.

[0096] S703 outputs the final matching table as the final allocation result. The final matching table records the unique on-site execution resource corresponding to each task.

[0097] The above process ensures optimal overall matching between tasks and on-site execution resources by setting the total fit score as the objective function and pursuing its maximization; it fundamentally ensures that the allocation results meet the actual on-site operational needs through dual constraints; and the final output matching table clearly records the unique correspondence between each task and its corresponding execution resource, providing a direct basis for subsequent instruction issuance and equipment execution, and reducing ambiguity in the scheduling process.

[0098] In one embodiment of the present invention, the following instruction issuance step is further included:

[0099] S801 reads the final allocation result and generates a scheduling instruction package for each pair of tasks and on-site execution resources. The instruction package fields include a task identifier, a resource identifier, starting position coordinates, and target position coordinates. The task identifier uniquely identifies the task to be executed, the resource identifier precisely locates the on-site execution resource to perform the task, the starting position coordinates mark the current location of the on-site execution resource, and the target position coordinates indicate the destination the task needs to reach. These fields ensure the uniqueness and directionality of the instructions, providing clear action guidelines for the on-site execution resources.

[0100] S802 sends the scheduling instruction packet to the corresponding field execution resource and waits for the field execution resource to return a receipt confirmation message. This process realizes closed-loop tracking of instruction transmission through a two-way communication mechanism, avoiding the problem of the instruction being in a black box state after it is issued, and ensuring that the system can grasp the instruction reception status in real time.

[0101] S803, after receiving the confirmation message, records the instruction status as sent, indicating that the instruction has been successfully transmitted to the execution end; if no confirmation is received within the preset timeout period, it is marked as pending retry and sent again after a fixed interval.

[0102] Through the above process, the instruction issuance step achieves a seamless connection from optimization results to on-site execution. The standardized design of instruction packet fields ensures that on-site execution resources can accurately understand task requirements and reduce execution deviations. The two-way communication and status tracking mechanism realizes the traceability of instruction transmission.

[0103] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0104] The embodiments of the present invention have been described above. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of the present embodiments, all of which are within the protection scope of the present embodiments.

Claims

1. A dynamic resource priority intelligent allocation method based on a device dispatching system, characterized in that, Includes the following steps: S101, collects device status data and resource status data; The on-site execution resources consist of location information and personnel information composed of equipment status data and resource status data, and each on-site execution resource corresponds to a unique location coordinate. S102, classify tasks based on task type and deadline, compare the remaining time of a task with the historical average remaining time of the same task type, and calculate the task urgency by combining the preset penalty coefficient. S103 divides the field execution resources into equipment layer, storage layer and human-machine collaboration layer, and further subdivides them according to physical lanes. Combining the idleness of each field execution resource with the moving average occupancy rate within the sliding time window, the resource heat index is calculated. S104, based on the voxel mesh structure preset in the three-dimensional space of the warehouse, marks the voxel units in it, and calculates the spatiotemporal potential energy value based on the shortest feasible path distance between the voxel where the target location of the task is located and the voxel where the on-site execution resource is located, the resource heat index, and the current cumulative waiting time of the task. The process of obtaining the spatiotemporal potential energy value includes: S401, based on the voxel where the task is located and the voxel where the on-site execution resources are located, uses the Dijkstra algorithm to obtain the shortest feasible path distance; S402, by calculating the difference between the current time and the task creation time, the current cumulative waiting time of the task is obtained; S403, calculate the average distance of the shortest feasible path in the past thirty days as the distance baseline value, and calculate the average current cumulative waiting time of the task in the past thirty days as the time baseline value; S404, multiply the ratio of the shortest feasible path distance to the distance benchmark value by the first preset weight to obtain the distance term, multiply the resource heat index by the second preset weight to obtain the resource heat term, multiply the ratio of the current cumulative waiting time to the time benchmark value by the third preset weight to obtain the time term, and sum the distance term and the resource heat term and then subtract the time term to obtain the spatiotemporal potential energy value. The values ​​of the first preset weight, the second preset weight and the third preset weight are in the range of 0 to 1, and their sum is 1. S105, construct a suitability matrix between the task and on-site execution resources based on task urgency, resource heat index, and spatiotemporal potential energy value, including: S501 linearly standardizes the task urgency and resource heat index to obtain standardized urgency and standardized heat. S502, calculate the ratio of the sum of the standardized urgency and the standardized popularity of OnePlus to obtain the benchmark matching value; S503: For each pair of tasks and on-site execution resources, the spatiotemporal potential energy value is taken as the inverse number and used as the exponent to obtain the exponential decay term. The exponential decay term is then multiplied by the benchmark matching value to obtain the fit score. S504, write the fit score into the position corresponding to the task and the on-site execution resources to complete the fit matrix filling; S106, perform optimization calculations to maximize the total fitness score of the fitness matrix, and obtain the final allocation result of the task and on-site execution resources.

2. The method of claim 1, wherein, Equipment status data includes: equipment location, idle time, failure probability, and task queue length; resource status data includes: task information, storage location information, and personnel information; task information includes: task type and task deadline. The calculation process for task urgency includes: S201, extract the task type, task deadline, and current time, establish a task type index, and determine the time base of the task; S202, take the difference between the task deadline and the current time as the remaining time of the task, and obtain the historical average remaining time under the same task type; S203, calculate the ratio of the remaining time of the task to the historical average remaining time to obtain the time pressure comparison coefficient; S204. Multiply the preset penalty coefficient by the time pressure comparison coefficient to obtain the task urgency.

3. The method of claim 1, wherein, The on-site execution resources are divided into equipment layer, storage layer and human-machine collaboration layer according to their functions, and the resource units are further subdivided within each layer according to the physical roadway number; Within a set sliding window, the occupancy status of resource units is continuously recorded to obtain the occupancy duration. The ratio of the occupancy duration to the duration of the sliding window is calculated to obtain the moving average occupancy rate of the resource unit. The resource heat index is calculated by weighting and integrating the difference between the availability and the occupancy rate with the moving average occupancy rate according to preset weights.

4. The dynamic resource priority intelligent allocation method based on an equipment scheduling system according to claim 1, characterized in that, Based on a pre-defined voxel mesh structure in the warehouse's 3D space, the voxel units within it are labeled, including: S301, the warehouse three-dimensional space is rasterized according to the preset voxel size, and a continuously numbered voxel index is generated; S302, Read the location coordinates of the on-site execution resource and map the on-site execution resource to the corresponding voxel index; S303, mark each voxel unit with a pass tag. The pass tags include: fixed obstacle tag, temporary occupation tag and passable tag.

5. The dynamic resource priority intelligent allocation method based on an equipment scheduling system according to claim 1, characterized in that, Optimization calculations are performed to maximize the total fitness score of the fitness matrix, resulting in the final allocation of tasks and on-site execution resources, including: S601, Read the fitness matrix, sum the fitness scores of all matrix elements, and set it as the objective function; S602 sets task uniqueness constraints and resource capacity constraints to limit the space of feasible matching solutions; S603 performs optimization operations under the objective function and constraints to obtain the final allocation result.

6. The dynamic resource priority intelligent allocation method based on an equipment scheduling system according to claim 1, characterized in that, The optimization calculations include: S701, traverse from high to low according to the suitability score, select on-site execution resources with available capacity for unassigned tasks, and generate an initial matching table; S702, perform on-site resource exchange within the initial matching table. If the total matching score increases after the exchange, retain the exchange and continue iterating until the total matching score no longer increases after two consecutive iterations. S703 outputs the final matching table as the final allocation result. The final matching table records the unique on-site execution resource corresponding to each task.

7. The dynamic resource priority intelligent allocation method based on an equipment scheduling system according to claim 1, characterized in that, It also includes the following instruction issuance steps: S801, read the final allocation result, generate a scheduling instruction package for each pair of tasks and on-site execution resources. The instruction package fields include task identifier, resource identifier, starting position coordinates and target position coordinates; S802, the scheduling instruction packet is sent to the corresponding field execution resource, and the field execution resource is waited for the return of the receipt confirmation information; S803 records the command status as sent after receiving the confirmation message. If no confirmation is received within the preset timeout period, it is marked as pending retry.

Images