Intelligent unmanned cabinet distributed replenishment task dynamic scheduling system

Abstract

The application relates to an intelligent unmanned cabinet distributed replenishment task dynamic scheduling system, belonging to the technical field of unmanned retail and intelligent scheduling, which comprises the following: a shortage state monitoring and forecasting module, which is used for acquiring inventory data of each cabinet and each channel in real time; a replenishment task generation module, which is used for generating replenishment tasks according to a replenishment task candidate set and a preset task generation strategy; a distributed scheduling engine, which is used for clustering and partitioning replenishment tasks according to geographical regions; a replenishment resource management module; a dynamic adjustment module; a task execution monitoring and feedback module. The intelligent unmanned cabinet distributed replenishment task dynamic scheduling system can predict the shortage time of each cabinet in advance based on a deep learning-based sales forecasting model through the shortage state monitoring and forecasting module, and automatically generates replenishment tasks before goods are sold out, thereby changing the traditional passive inspection mode into an active predictive replenishment mode; compared with the traditional fixed time interval inspection mode, the shortage rate is significantly reduced, and the availability of goods is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of unmanned retail and intelligent scheduling technology, specifically to an intelligent unmanned locker distributed replenishment task dynamic scheduling system, which is particularly suitable for the intelligent generation, dynamic allocation and path optimization of replenishment tasks in large-scale unmanned retail locker networks. Background Technology

[0002] With the rapid development of the unmanned retail industry, large chain unmanned retail enterprises have deployed thousands to tens of thousands of unmanned retail kiosks, widely distributed in various locations such as office buildings, residential communities, transportation hubs, and schools. The replenishment operation of unmanned retail kiosks faces increasingly severe challenges: a wide variety of products (tens to hundreds of SKUs per kiosk), frequent sales fluctuations, and dynamic changes in stockout status across kiosks. How to efficiently allocate replenishment tasks to limited replenishment resources (replenishment personnel or a fleet of replenishment robots) while minimizing operating costs while meeting timeliness requirements has become a core problem restricting the large-scale operation of unmanned retail.

[0003] The existing unmanned retail locker replenishment scheduling technology has the following main shortcomings: First, replenishment task generation is passively delayed. Existing systems mostly use fixed-interval inspections or manual visual checks to detect stockouts. The generation of replenishment tasks lags far behind actual demand, making it impossible to proactively trigger replenishment tasks when goods are about to sell out, resulting in stockout losses.

[0004] Secondly, task allocation relies on human experience and lacks global optimization. After replenishment tasks are generated, they are mostly manually assigned to replenishment personnel or replenishment robots by schedulers based on their experience, without considering factors such as the current location, remaining capacity, skill matching degree, and historical efficiency of each replenishment resource. This results in unreasonable task allocation, low utilization rate of replenishment resources, and high empty running rate.

[0005] Third, the path planning is static and rigid. After assigning replenishment tasks, the existing system mostly adopts fixed routes or simple proximity principles for replenishment paths, without comprehensively considering dynamic factors such as traffic conditions, the urgency of replenishment for each container, and the differences in replenishment time. This results in low replenishment efficiency and the inability to prioritize urgent stockouts.

[0006] Fourth, the system lacks the ability to respond to sudden demand. When sudden stockouts occur (such as a surge in sales at a particular counter due to surrounding events) or replenishment resource failures (such as a replenishment robot temporarily going offline), the existing system cannot detect and dynamically adjust the task allocation plan in real time, resulting in a slow emergency response.

[0007] Fifth, it lacks distributed collaborative scheduling capabilities. As the number of unmanned retail kiosks grows to several thousand, the scale of the replenishment scheduling problem increases dramatically. The computational load of traditional centralized scheduling methods explodes, making it impossible to provide a feasible solution within an acceptable timeframe.

[0008] Therefore, there is an urgent need for a distributed scheduling system capable of intelligently generating, dynamically allocating, and collaboratively optimizing replenishment tasks. This invention aims to provide a distributed dynamic scheduling system for replenishment tasks in intelligent unmanned vending machines, addressing the problems of delayed task generation, reliance on manual allocation, rigid path planning, and lack of distributed collaborative capabilities in existing replenishment scheduling technologies. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention provides a dynamic scheduling system for distributed replenishment tasks in intelligent unmanned vending machines. It has advantages such as proactive predictive task generation, distributed partitioned parallel scheduling, dynamic incremental rescheduling, and multi-objective path optimization. It solves the problems of delayed task generation, lack of global optimization in manual allocation, inability of static path planning to cope with dynamic changes, and inability of centralized scheduling to support large-scale networks in existing replenishment scheduling.

[0010] To achieve the above objectives, the present invention provides the following technical solution: A dynamic scheduling system for distributed replenishment tasks of intelligent unmanned vending machines includes: The out-of-stock status monitoring and prediction module is connected to the inventory perception system of each smart unmanned cabinet. It is used to obtain the inventory data of each cabinet and each channel in real time, predict the out-of-stock time point of each cabinet in the future time window based on the inventory change trend, and generate a candidate set of replenishment tasks including replenishment priority, expected out-of-stock time and suggested replenishment quantity. The replenishment task generation module, connected to the stockout status monitoring and prediction module, is used to generate replenishment tasks containing task number, target cabinet number, replenishment product list, suggested replenishment quantity, expected completion time, and task urgency level based on the replenishment task candidate set and preset task generation strategy. A distributed scheduling engine, connected to the replenishment task generation module, is used to cluster and partition replenishment tasks by geographical region, perform distributed matching and path optimization of replenishment tasks and available replenishment resources in each region, and generate task allocation schemes and optimal replenishment paths for each replenishment resource. The replenishment resource management module is used to maintain the operational status information of each replenishment resource in real time. The operational status information includes the current location, available capacity, remaining range, current task execution status, and historical efficiency indicators. The dynamic adjustment module is connected to the distributed scheduling engine and the replenishment resource management module respectively. When a new emergency replenishment task or replenishment resource anomaly is detected, it triggers incremental rescheduling of the replenishment task allocation scheme in the affected area and outputs the adjusted task allocation scheme. The task execution monitoring and feedback module is used to track the task execution progress of each replenishment resource in real time, and feed back the execution completion status and actual time consumption to the distributed scheduling engine for subsequent scheduling strategy optimization.

[0011] Furthermore, the out-of-stock status monitoring and prediction module adopts a sales forecasting model based on long short-term memory networks. It predicts the sales volume of each product in the next 24 hours based on historical sales data and influencing factors, and calculates the expected out-of-stock time in combination with the current inventory.

[0012] Furthermore, the distributed scheduling engine clusters replenishment tasks by geographical region and uses clustering algorithms based on K-means or DBSCAN to aggregate spatially adjacent replenishment tasks into several task clusters. Each task cluster is independently scheduled by the replenishment resources in that region, realizing distributed parallel solution of large-scale problems.

[0013] Furthermore, the distributed scheduling engine employs a multi-objective optimization algorithm for task allocation within each region, with the comprehensive optimization objectives being to minimize the total replenishment completion time, minimize the total travel distance of replenishment resources, and maximize the priority fulfillment of urgent out-of-stock tasks.

[0014] Furthermore, the incremental rescheduling strategy of the dynamic adjustment module is as follows: only the regions where new replenishment tasks are located or the regions affected by abnormal replenishment resources are rescheduled, while the scheduling schemes for other regions remain unchanged.

[0015] Furthermore, the task urgency level in the replenishment task generation module includes at least three levels: urgent, regular, and planned. The expected completion time of the urgent task is less than a first preset threshold, the expected completion time of the regular task is less than a second preset threshold, and the expected completion time of the planned task is less than a third preset threshold.

[0016] Furthermore, the replenishment task generation module also includes a task merging strategy, which automatically merges multiple replenishment tasks into a single comprehensive replenishment task when multiple replenishment tasks exist for the same target container within a preset time window.

[0017] Furthermore, the system also includes a visual scheduling dashboard module, which displays the real-time location, task allocation status, replenishment path, and task execution progress of each replenishment resource.

[0018] Furthermore, the replenishment resources include one or more of replenishment personnel and replenishment robots.

[0019] Furthermore, the replenishment resource management module calculates the average replenishment time per cabinet, average travel speed, on-time task completion rate, and overall efficiency score for each replenishment resource, which is used by the distributed scheduling engine for resource matching when allocating tasks.

[0020] Compared with existing technologies, the present invention provides a dynamic scheduling system for distributed replenishment tasks of intelligent unmanned lockers, which has the following beneficial effects: 1. This intelligent unmanned locker distributed replenishment task dynamic scheduling system, through the out-of-stock status monitoring and prediction module, uses a deep learning sales prediction model to predict the out-of-stock time of each locker in advance, and actively generates replenishment tasks before the goods are sold out, transforming the traditional passive inspection mode into an active predictive replenishment mode; compared with the traditional fixed time interval inspection method, the out-of-stock rate is significantly reduced and the availability of goods is greatly improved.

[0021] 2. This intelligent unmanned locker distributed replenishment task dynamic scheduling system uses a distributed scheduling engine to cluster and partition large-scale replenishment tasks by geographical region and then solve them in parallel. This overcomes the problem of explosive growth in computational load when the task scale increases, which is a problem of traditional centralized scheduling methods. Each region is independently scheduled and computed in parallel. The task allocation scheme can be generated within an acceptable time and supports the smooth expansion of the unmanned retail locker network from hundreds to tens of thousands of units.

[0022] 3. This intelligent unmanned locker distributed replenishment task dynamic scheduling system, through the incremental rescheduling strategy of the dynamic adjustment module, only performs local rescheduling on the affected area when new emergency tasks are added or replenishment resources are abnormal, without the need for global recalculation. This not only ensures the system's rapid response to sudden changes, but also maintains the stability of the scheduling scheme in most areas, significantly improving the system's robustness in dynamic operating environments. Attached Figure Description

[0023] Figure 1 This is an overall system block diagram of a distributed replenishment task dynamic scheduling system for intelligent unmanned lockers according to the present invention.

[0024] Figure 2 This invention presents a flowchart of the distributed scheduling engine of a distributed replenishment task dynamic scheduling system for intelligent unmanned vending machines.

[0025] Figure 3 This is a flowchart of the dynamic incremental rescheduling process of a distributed replenishment task dynamic scheduling system for intelligent unmanned lockers according to the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Please see Figures 1 to 3 .

[0028] Example 1: System Overall Architecture and Module Deployment This embodiment provides a dynamic scheduling system for distributed replenishment tasks of intelligent unmanned vending machines. It is deployed on the intelligent operation platform of a large unmanned retail enterprise to manage 2,000 unmanned retail vending machines and 50 replenishment resources (including 35 replenishment personnel and 15 replenishment robots) distributed in various areas of the city.

[0029] Out-of-stock monitoring and prediction module: Each unmanned retail cabinet is equipped with a weight sensor and a photoelectric sensor, reporting real-time inventory data for each aisle every 5 minutes. The module uses an LSTM-based sales forecasting model (inputs are sales data for each time period over the past 7 days, weather data, holiday data, and surrounding event data; output is the predicted sales volume for each hour over the next 24 hours), combined with the current inventory level, to calculate the estimated out-of-stock time for each product at each point in time. When the estimated out-of-stock time for a product is less than a preset threshold (e.g., 6 hours), a replenishment task candidate is automatically generated. This module also supports emergency triggering based on real-time out-of-stock status (replenishment candidates are immediately generated when the inventory in a certain aisle falls below the safety stock line).

[0030] Replenishment Task Generation Module: After receiving the candidate set of replenishment tasks, it generates standardized replenishment tasks according to the preset task generation strategy. The task includes: task number, target cabinet number, replenishment product list (product name + suggested replenishment quantity), expected completion time (based on the out-of-stock prediction time 2 hours in advance), and task urgency level (urgent level: out-of-stock time < 2 hours; normal level: out-of-stock time 2-8 hours; planned level: out-of-stock time 8-24 hours).

[0031] Distributed scheduling engine: Deployed on a cloud server cluster, it adopts a divide-and-conquer strategy to achieve distributed scheduling. First, it divides the region based on the geographical location and number of each replenishment task (in this embodiment, it is divided into 12 regions). The task allocation and path optimization in each region are solved in parallel by an independent scheduling thread.

[0032] Replenishment Resource Management Module: The module obtains the location of each resource in real time through the GPS positioning of the replenishment personnel's mobile APP and the vehicle positioning system of the replenishment robot; it obtains the current task status (idle / running / faulty / offline), remaining available capacity and remaining range of each resource through the task management system; and it calculates the average replenishment efficiency of each resource through historical task data.

[0033] Dynamic adjustment module: Continuously monitors new tasks and resource status changes, triggering incremental rescheduling.

[0034] Task execution monitoring and feedback module: Track task progress in real time through replenishment personnel's APP check-in and replenishment robot task logs.

[0035] Example 2: Detailed Workflow of the Distributed Scheduling Engine This embodiment details the workflow of the distributed scheduling engine.

[0036] Step 1: Task Clustering Partitioning The distributed scheduling engine receives a list of replenishment tasks from the replenishment task generation module (assuming there are currently 150 replenishment tasks to be assigned, distributed across various areas of the city). The scheduling engine employs a clustering strategy based on geographical location and the number of tasks. First, the tasks are clustered using DBSCAN based on geographical coordinates (cluster radius ε = 3km, minimum number of tasks within a cluster minPts = 5), resulting in 150 tasks clustered into several natural clusters. Then, regions are merged or split based on the number of tasks in each cluster and the amount of available replenishment resources, ensuring that the number of tasks in each region matches the amount of available replenishment resources (ideally, each replenishment resource should handle 3-8 tasks), ultimately forming 12 scheduling regions. Each region has 8-18 replenishment tasks, allocated 2-5 replenishment resources.

[0037] Step 2: Task allocation and route optimization within the region The replenishment task allocation problem for each region is modeled as a vehicle routing problem with time window constraints. For a given region (assuming it contains 10 replenishment tasks and 2 replenishment resources), the scheduling engine uses an improved genetic algorithm to solve the problem, with the comprehensive optimization objectives of minimizing total completion time, minimizing total travel distance, and maximizing the priority of urgent tasks. Encoding scheme: Integer encoding is used, with each chromosome representing the task allocation order and resource allocation.

[0038] Initial population generation: Initial solutions are generated based on a greedy strategy (nearest allocation + priority of urgent needs).

[0039] Genetic operations: The optimal solution is output after 500 iterations using ordered crossover and exchange mutation operators.

[0040] Step 3: Optimal replenishment route generation For each replenishment resource, an optimal driving route is generated based on its assigned replenishment task and task execution order. The route takes into account the actual geographical location of each unmanned retail cabinet (based on real-time traffic conditions from the Gaode Map API), the expected completion time window of each task, the average replenishment time of each cabinet (based on historical statistics), and the available capacity limit of the replenishment resource.

[0041] Parallel solution: The scheduling tasks of 12 regions are solved simultaneously by 12 parallel threads. Each region is optimized independently, and the total solution time is about 12 seconds, which is far better than centralized scheduling (centralized solution of the same scale tasks takes about 180 seconds).

[0042] Example 3: Specific Implementation of Dynamic Incremental Rescheduling This embodiment details the incremental rescheduling process of the dynamic adjustment module in the event of an emergency.

[0043] Scenario: The system completed a full scheduling round at 10:00, generating an allocation plan covering 150 tasks. At 10:15, a community unmanned retail cabinet experienced a surge in sales of beverages due to a sudden promotional event in the surrounding area, triggering an emergency restocking task (task urgency level: emergency, expected completion time: <1 hour).

[0044] Step 1, Anomaly Detection and Triggering: The dynamic adjustment module detects a new emergency replenishment task, locates its scheduling region (assuming it is region 3) based on the task's geographical location, and triggers incremental rescheduling of region 3.

[0045] Step 2: Determine the scope of impact: Determine the potential impact of the new task on the existing tasks in Region 3 (a total of 12 tasks, including 3 replenishment resources) and the new task. Merge the new task with the existing tasks in Region 3, while keeping the scheduling schemes for the other 11 regions unchanged.

[0046] Step 3, Incremental Rescheduling: Reschedule the 13 tasks in Region 3 (12 original tasks + 1 new task). Unlike full rescheduling, incremental rescheduling retains the allocation of tasks that have already started executing (which cannot be changed), and only re-optimizes the allocation of tasks that have not yet started executing and the new task.

[0047] Step 4: Adjustment Scheme Generation: The incremental rescheduling algorithm (based on an improved genetic algorithm, with the number of iterations reduced to 100 generations to ensure rapid response) generates a new allocation scheme within 8 seconds: the emergency task is assigned to the nearest idle replenishment robot (approximately 1.2km), which is expected to arrive and complete the replenishment within 20 minutes; a regular task originally assigned to this robot is transferred to another replenishment personnel in the same area, with the impact limited to area 3.

[0048] Step 5: Plan Distribution and Execution: The adjusted task allocation plan is distributed to relevant replenishment resources via APP push and robot control system, and replenishment personnel in other areas are not affected in any way.

[0049] In contrast, global rescheduling would require approximately 180 seconds to recalculate all 150 tasks, and would cause frequent changes to the task plans already issued to the other 11 regions, resulting in execution chaos. Incremental rescheduling completes in just 8 seconds and affects only one region, demonstrating the robustness and rapid response advantages of this invention.

[0050] Example 4: Replenishment Resource Management and Efficiency Evaluation This embodiment details the real-time monitoring and efficiency evaluation mechanism of replenishment resources by the replenishment resource management module.

[0051] Real-time status monitoring: The system obtains the status information of each replenishment resource in real time through the following methods: Location information: Replenishment personnel use GPS positioning via a mobile app (reporting their location every 30 seconds), while replenishment robots use an onboard positioning system (reporting their location every 10 seconds).

[0052] Task status: Replenishment personnel update their status by checking in via the "Start Task / Complete Task" function in the app; the replenishment robot automatically reports the task execution progress (arrived / replenishing / completed / abnormal).

[0053] Available capacity: The replenishment robot automatically reports the remaining load and remaining space; replenishment personnel manually enter the remaining available space through the APP.

[0054] Battery life: The replenishment robot automatically reports the remaining battery power / battery range; replenishment personnel record the fuel / battery status in the APP.

[0055] Efficiency evaluation metrics: Based on historical data, the system calculates the following efficiency metrics for each replenishment resource: Average replenishment time per container (by product type); Average driving speed (by time of day / by area); On-time completion rate of tasks; Quantity of goods replenished per unit time; Overall efficiency score (weighted average).

[0056] Resource matching in scheduling decisions: When allocating tasks, the distributed scheduling engine considers not only distance factors but also the efficiency indicators of each replenishment resource for optimal matching—high-urgency tasks are preferentially assigned to replenishment resources with high historical efficiency and on-time rate; replenishment tasks for specific product types are preferentially assigned to resources with rich experience in replenishing that type of product.

[0057] Example 5: Verification of Large-Scale Operational Effectiveness This embodiment uses actual operational data from a large unmanned retail enterprise to verify the effectiveness of the present invention.

[0058] Experimental setup: The company operates 2000 unmanned retail kiosks, with a daily replenishment task of approximately 800-1200 units and 50 replenishment resources. The operational data before and after the system's implementation is compared.

[0059] The experimental results are shown in Table 1.

[0060] Table 1: Conclusion: The system of the present invention is significantly superior to the traditional manual scheduling method in key indicators such as task allocation efficiency, response speed, resource utilization and operational quality.

[0061] Comparative Example 1: Centralized Scheduling Method A centralized scheduling method is adopted to unify the replenishment tasks of 2,000 cabinets into a central node for global optimization and solution. No regional clustering or partitioning is performed, and the other conditions are the same as those in this invention.

[0062] Comparison results: When the number of replenishment tasks exceeds 500, the solution time of the centralized scheduling algorithm increases exponentially (exceeding 180 seconds), failing to meet the requirements of real-time scheduling; when the number of replenishment tasks exceeds 1000, the solution time exceeds 10 minutes, rendering it essentially unusable. The distributed scheduling method of this invention solves the problem in parallel through regional partitioning, requiring only 12 seconds for 150 tasks, consistently maintaining good real-time performance.

[0063] Comparative Example 2: Static Fixed Route Scheduling A static fixed-route scheduling method is adopted—each replenishment resource replenishes goods according to a pre-defined responsibility area and fixed route, without dynamic task allocation or real-time path optimization.

[0064] Comparison results: Under the static fixed-route method, the stockout rate fluctuates significantly with sales volume (when sales volume increases by 30%, the stockout rate rises from 6.5% to 12.3%), failing to adapt to dynamic changes. The dynamic scheduling method of this invention can still stably control the stockout rate between 2.0% and 3.5% within a sales volume fluctuation range of ±50%, demonstrating excellent adaptability.

[0065] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0066] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A dynamic scheduling system for distributed replenishment tasks of intelligent unmanned vending machines, characterized in that, include: The out-of-stock status monitoring and prediction module is connected to the inventory perception system of each smart unmanned cabinet. It is used to obtain the inventory data of each cabinet and each channel in real time, predict the out-of-stock time point of each cabinet in the future time window based on the inventory change trend, and generate a candidate set of replenishment tasks including replenishment priority, expected out-of-stock time and suggested replenishment quantity. The replenishment task generation module, connected to the stockout status monitoring and prediction module, is used to generate replenishment tasks containing task number, target cabinet number, replenishment product list, suggested replenishment quantity, expected completion time, and task urgency level based on the replenishment task candidate set and preset task generation strategy. A distributed scheduling engine, connected to the replenishment task generation module, is used to cluster and partition replenishment tasks by geographical region, perform distributed matching and path optimization of replenishment tasks and available replenishment resources in each region, and generate task allocation schemes and optimal replenishment paths for each replenishment resource. The replenishment resource management module is used to maintain the operational status information of each replenishment resource in real time. The operational status information includes the current location, available capacity, remaining range, current task execution status, and historical efficiency indicators. The dynamic adjustment module is connected to the distributed scheduling engine and the replenishment resource management module respectively. When a new emergency replenishment task or replenishment resource anomaly is detected, it triggers incremental rescheduling of the replenishment task allocation scheme in the affected area and outputs the adjusted task allocation scheme. The task execution monitoring and feedback module is used to track the task execution progress of each replenishment resource in real time, and feed back the execution completion status and actual time consumption to the distributed scheduling engine for subsequent scheduling strategy optimization.

2. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The out-of-stock status monitoring and prediction module adopts a sales forecasting model based on long short-term memory networks. It predicts the sales volume of each product in the next 24 hours based on historical sales data and influencing factors, and calculates the expected out-of-stock time in combination with the current inventory.

3. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The distributed scheduling engine clusters replenishment tasks by geographical region and uses clustering algorithms based on K-means or DBSCAN to aggregate spatially adjacent replenishment tasks into several task clusters. Each task cluster is independently scheduled by the replenishment resources in that region, realizing distributed parallel solution of large-scale problems.

4. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The distributed scheduling engine employs a multi-objective optimization algorithm for task allocation within each region, with the comprehensive optimization objectives being to minimize the total replenishment completion time, minimize the total travel distance of replenishment resources, and maximize the priority fulfillment of urgent out-of-stock tasks.

5. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The incremental rescheduling strategy of the dynamic adjustment module is as follows: only the regions where new replenishment tasks are located or the regions affected by abnormal replenishment resources are rescheduled, while the scheduling schemes for other regions remain unchanged.

6. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The task urgency level in the replenishment task generation module includes at least three levels: urgent, regular, and planned. The expected completion time of the urgent task is less than a first preset threshold, the expected completion time of the regular task is less than a second preset threshold, and the expected completion time of the planned task is less than a third preset threshold.

7. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The replenishment task generation module also includes a task merging strategy, which automatically merges multiple replenishment tasks into a single comprehensive replenishment task when multiple replenishment tasks exist for the same target container within a preset time window.

8. The intelligent unmanned locker distributed replenishment task dynamic scheduling system according to claim 1, characterized in that, The system also includes a visual scheduling dashboard module, which displays the real-time location, task allocation status, replenishment path, and task execution progress of each replenishment resource.

9. A dynamic scheduling system for distributed replenishment tasks of an intelligent unmanned vending machine according to claim 1, characterized in that, The replenishment resources include one or more of replenishment personnel and replenishment robots.

10. A dynamic scheduling system for distributed replenishment tasks of an intelligent unmanned vending machine according to claim 1, characterized in that, The replenishment resource management module calculates the average replenishment time per cabinet, average travel speed, on-time task completion rate, and overall efficiency score for each replenishment resource, which is used by the distributed scheduling engine for resource matching when allocating tasks.

Images