Method and system for dynamic scheduling of mobile device resources in a smart park

Abstract

The application provides a kind of wisdom park mobile device resource dynamic scheduling method and system, belong to resource scheduling technical field, the wisdom park mobile device resource dynamic scheduling method of the present application includes: management record has multiple charging resources space-time map, the space-time map includes each charging resource in future time axis Multiple available time periods;When target mobile device is in schedulable state, determine the next task to be executed by the target mobile device, and estimate the energy consumption required to complete the next task;In the case where the result of the current power of the target mobile device minus the estimated energy consumption is less than the preset power threshold, select a target charging resource for the target mobile device based on the space-time map.The method can improve the global application efficiency of park resources.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of resource scheduling, and in particular to a dynamic scheduling method and system for mobile devices in a smart park. BACKGROUND

[0002] In modern smart factories and logistics parks, devices such as automated guided vehicles (AGVs) or autonomous mobile robots (AMRs) (hereinafter collectively referred to as "mobile devices") undertake key material handling tasks. These devices are battery-powered and need to return to designated charging piles for charging regularly. In this case, the number of charging piles is limited, and during peak hours, multiple mobile devices may need to be charged at the same time, resulting in queuing and waiting, which reduces the utilization rate of the devices. It is necessary to consider scheduling the devices to charge intelligently under the premise of ensuring that the handling tasks are completed in time, so as to avoid interrupting the tasks due to low battery or delaying the tasks due to unnecessary charging.

[0003] Traditional scheduling systems are usually based on fixed thresholds (e.g., charging when the battery level is below 20%) or simple queuing logic, and lack prediction of future tasks and charging pile usage, resulting in suboptimal scheduling decisions and easy congestion or resource idling. Most existing technical solutions focus on complex path planning algorithms or intelligent scheduling algorithms based on reinforcement learning. These solutions are complex to implement, have high computational costs, and are prone to local optima, requiring precise environmental models and being difficult to quickly deploy and adjust in actual production environments. Therefore, there is an urgent need for a way to improve global scheduling effectiveness to solve the above problems. SUMMARY

[0004] The present application provides a dynamic scheduling method and system for mobile devices in a smart park to solve the problem of poor global scheduling effectiveness in the prior art and achieve the effect of globally optimized scheduling.

[0005] The present application provides a dynamic scheduling method for mobile devices in a smart park, which is applied to a management scheduling party, and the method comprises:

[0006] A time-space map recording a plurality of charging resources is managed, and the time-space map includes a plurality of available time periods of each charging resource on a future time axis;

[0007] When the target mobile device is in a schedulable state, the next task to be performed by the target mobile device is determined, and the energy consumption required to complete the next task is estimated;

[0008] In a case that a result of the current power of the target mobile device minus the estimated energy consumption is less than a preset power threshold, a target charging resource is selected for the target mobile device based on the space-time map; after the target mobile device completes charging at the target charging resource in the target time period, the target mobile device travels from the target charging resource to a high-frequency task area corresponding to the target time period, and an opportunity cost index required by the target mobile device is the lowest.

[0009] According to the application, a dynamic scheduling method of mobile device resources in a smart park is provided, and the method is also applied to allocation decision of a new task, which comprises the following steps:

[0010] In the task allocation, an estimated opportunity cost index that each candidate mobile device can obtain after executing the new task is determined based on the space-time map.

[0011] The candidate mobile device with the lowest estimated opportunity cost index is selected as the final execution mobile device of the new task.

[0012] According to the application, a dynamic scheduling method of mobile device resources in a smart park is provided, and the method is also applied to allocation decision of a new task, which comprises the following steps:

[0013] A time axis of each charging resource is divided into different time period layers, the time period layers comprise a near view layer, a middle view layer and a far view layer, the near view layer is used for locking a reserved charging resource in a first time period by a hard reservation mark; the middle view layer is used for publishing a predicted available window in a second time period to determine the target charging resource; and the far view layer is used for marking a maintenance or failure plan in a third time period as a prerequisite exclusion condition for decision making.

[0014] The length of the third time period is greater than the length of the second time period, and the length of the second time period is greater than the length of the first time period.

[0015] According to the application, a dynamic scheduling method of mobile device resources in a smart park is provided, and the method is also applied to allocation decision of a new task, which comprises the following steps:

[0016] All predicted available windows that are not marked as unavailable by the far view layer and have a sufficient length are selected as candidate windows in a candidate set in the middle view layer.

[0017] When each candidate window in the candidate set is evaluated, an opportunity cost index of each candidate window is calculated.

[0018] A charging resource corresponding to a candidate window with the lowest opportunity cost index is determined as the target charging resource, and a target time period is determined.

[0019] According to the intelligent park mobile device resource dynamic scheduling method provided by the application, the opportunity cost index is determined based on at least one of the following: the predicted driving time of the target mobile device to the charging resource corresponding to the candidate window; the predicted duration of completing charging in the candidate window; and the predicted driving time from the charging resource to the high-frequency task area after completing charging.

[0020] According to the intelligent park mobile device resource dynamic scheduling method provided by the application, the method further comprises: applying a time-sensitive soft reservation to the candidate window being evaluated in the process of calculating and comparing the opportunity cost index; and upgrading the soft reservation to a hard reservation located in the near view layer after determining the target charging resource, and releasing the soft reservation of other charging resources that are not selected.

[0021] According to the intelligent park mobile device resource dynamic scheduling method provided by the application, the method further comprises: in the case that a mobile device completes charging in advance and releases the charging resource, clearing the hard reservation mark of the released charging resource in the near view layer, and pushing the updated idle period of the released charging resource to a device queue that is waiting for charging or the opportunity cost index calculation result does not meet the preset condition, as a potential immediately executable low-cost charging opportunity in the device queue.

[0022] According to the intelligent park mobile device resource dynamic scheduling method provided by the application, the high-frequency task area is a dynamic geographic fence, and the position and range of the high-frequency task area are predicted and updated according to the historical task heat map in the time span of the middle view layer.

[0023] According to the intelligent park mobile device resource dynamic scheduling method provided by the application, the size of the preset power threshold is inversely proportional to the number of charging resources corresponding to the predicted available window in the middle view layer.

[0024] The application further provides an intelligent park mobile device resource dynamic scheduling system, comprising:

[0025] The management module is configured to manage a space-time map recording a plurality of charging resources, wherein the space-time map comprises a plurality of available time periods of each charging resource on a future time axis.

[0026] The first processing module is configured to determine a next task to be executed by the target mobile device when the target mobile device is in a schedulable state, and estimate the energy consumption required for completing the next task.

[0027] The second processing module is configured to select a target charging resource for the target mobile device based on the space-time map when a result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold; and the target mobile device has the lowest opportunity cost index required for traveling from the target charging resource to a high-frequency task area corresponding to a target time period after completing charging at the target charging resource in the target time period.

[0028] The application further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the smart park mobile device resource dynamic scheduling method according to any one of the above when executing the program.

[0029] The application further provides a non-transitory computer readable storage medium, which stores a computer program, and the computer program is executable on the processor to implement the smart park mobile device resource dynamic scheduling method according to any one of the above.

[0030] The application further provides a computer program product, which includes a computer program, and the computer program is executable on the processor to implement the smart park mobile device resource dynamic scheduling method according to any one of the above.

[0031] The smart park mobile device resource dynamic scheduling method and system provided by the application take the computer opportunity cost index as the core, combine the dynamic high-frequency task area prediction and the hierarchical linkage of task allocation, and make the charging problem of the smart park mobile device in a complex dynamic environment be scheduled, so that each charging decision is no longer isolated, but serves the long goal of maximizing the whole life cycle operation efficiency of the device, and the real global optimization scheduling is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0033] Figure 1 is a flowchart of the smart park mobile device resource dynamic scheduling method provided by the application;

[0034] Figure 2 is a schematic diagram of the space-time map of the smart park mobile device resource dynamic scheduling method provided by the application;

[0035] Figure 3It is a structural schematic diagram of the intelligent park mobile device resource dynamic scheduling system provided by the application.

[0036] Figure 4 is a structural schematic diagram of an electronic device provided by the application. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0038] The intelligent park mobile device resource dynamic scheduling method and system of the present application will be described below in combination with Figures 1-4 the drawings.

[0039] The embodiments of the present application provide an intelligent park mobile device resource dynamic scheduling method and system, which are especially suitable for intelligent logistics parks, automated ports and other scenarios with a large number of autonomous mobile robots (AGV), automatic guided vehicles (AGC) or unmanned vehicles and other mobile devices. The core goal of the method is to efficiently complete charging and energy replenishment while performing business tasks, thereby improving the operation efficiency of the entire device cluster.

[0040] In view of the response mode of charging nearby when the power is low in the prior art, the present application proposes a scheduling idea oriented to the optimal opportunity cost. It not only finds a charging pile for the device, but also finds a correct charging pile at the correct time for the device, so that it can seamlessly enter the task with the highest future value in the best state after charging is completed, thereby realizing the optimization of global efficiency.

[0041] As shown in Figure 1 , the intelligent park mobile device resource dynamic scheduling method of the embodiments of the present application mainly includes steps 110, 120 and 130.

[0042] Step 110: managing a space-time map recording a plurality of charging resources, the space-time map including a plurality of available time periods of each charging resource on a future time axis;

[0043] Step 120: determining a next task to be executed by the target mobile device when the target mobile device is in a schedulable state, and estimating the energy consumption required to complete the next task;

[0044] Step 130: selecting a target charging resource for the target mobile device based on the space-time map in the case that the result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold.

[0045] The target mobile device has the lowest opportunity cost index for traveling from the target charging resource to the high-frequency task area corresponding to the target time period after completing charging at the target charging resource in the target time period.

[0046] The space-time map includes a basic data structure that not only records the map of the geographic location (spatial dimension) of the charging resources such as charging piles, but also dynamically manages the charging plan of each charging pile on its future time axis (temporal dimension). A plurality of available time periods are time periods marked as available for scheduling on the time axis, which can be divided into different time period layers according to their properties.

[0047] The target mobile device refers to any mobile device that needs to be evaluated whether it needs to be charged. The schedulable state refers to the state in which the device is currently idle or will soon complete the current task and can accept new instructions. Instead of passively waiting for the device to issue a charging request, the system actively predicts the future tasks of the device when it is schedulable. Determining the next task can be achieved by querying the task management system's task queue to be allocated, or predicting the highest probability of the next task based on historical data and the current location of the device. The system can also estimate the energy consumption by calculating the energy consumption model of different types of mobile devices based on parameters such as the path length of the task, the load weight, the device model, the battery state of health (SoH), etc.

[0048] In the case where the result of subtracting the estimated energy consumption from the current power of the target mobile device is less than the preset power threshold, the target charging resource is selected for the target mobile device based on the space-time map. Assuming that the device completes the next task, it can be determined how much power it will have left. Only when this predicted future power is lower than the preset power threshold, the charging scheduling process is started.

[0049] By way of example, the preset power threshold is dynamic rather than fixed, which makes the triggering time more intelligent. Once triggered, the system starts making decisions based on the space-time map.

[0050] The target charging resource and the target time period are the results of this scheduling decision. The high-frequency task area is not a fixed geographic area, but a region where the system dynamically determines that the tasks will be most intensive or the highest value in a certain time period in the future, i.e., the time period after the target device completes charging.

[0051] The target mobile device has the lowest opportunity cost index for traveling from the target charging resource to the high-frequency task area corresponding to the target time period after completing charging at the target charging resource in the target time period.

[0052] The opportunity cost index is a dimensionless scalar function for quantifying the system-level time cost of a mobile device after completing charging at a specified charging resource and re-engaging in a next high-value task. The smaller the opportunity cost index value, the lower the negative impact of the charging decision on the overall carrying efficiency in the time-space dimension, that is, the lower the opportunity cost.

[0053] The opportunity cost index is determined based on at least one of the following: a predicted travel time of the target mobile device to reach the charging resource corresponding to the candidate window; a predicted duration for completing charging in the candidate window; and a predicted travel time from the charging resource to the high-frequency task area after completing charging.

[0054] In calculating the opportunity cost, each factor can be weighted and summed, or a correction factor with corresponding influence can also be combined for calculation, which is not limited here.

[0055] The high-frequency task area is a dynamic geofence, and the location and range of the high-frequency task area are predicted and updated according to a historical task heat map in a time span of the middle scene layer.

[0056] The high-frequency task area is a dynamic geofence that rolls over time, and its location and range are automatically generated only by statistically analyzing the density peak of the historical task heat map in the middle scene layer and are updated offline once a day, without the need for additional algorithms to quickly lock the area where new tasks are most likely to occur in the next stage.

[0057] It can be understood that time can be used as the only common dimension to convert the entire process of travel-charging-retravel into a single index that can be sorted, avoiding multi-objective optimization.

[0058] In other words, unlike the prior art, in the present embodiment, the final evaluation criterion for selecting a charging resource is not which charging pile is closest to the current charging, but rather which charging pile, after charging is completed, can reach the next high-frequency task area at the lowest cost, thereby improving the utilization efficiency of the mobile device.

[0059] The travel time can be an implementation of the opportunity cost index, and the shorter the time, the lower the opportunity cost, and the faster the device can be engaged in a new round of efficient work.

[0060] The intelligent park mobile device resource dynamic scheduling system provided by the embodiment of the present application constructs a multi-layer space-time map, takes the opportunity cost index as the core, and combines the hierarchical linkage of dynamic high-frequency task area prediction and task allocation, so that the intelligent park mobile device can be scheduled to solve the charging problem in a complex dynamic environment. It makes each charging decision no longer isolated, but serves the long goal of maximizing the whole life cycle operation efficiency of the device, and realizes real global optimization scheduling.

[0061] In some implementations, to effectively support the calculation of the opportunity cost index described above, the present application provides a refined spatiotemporal map. As shown in FIG. 8, which shows a schematic diagram of a spatiotemporal map, the horizontal axis is used to represent time, and in the vertical direction, it is divided into three logical time period levels. Figure 2

[0062] Near view layer C1 : can manage the deterministic state within 0-30 minutes. As shown in the figure, AGV-007 has a hard appointment at the current 9:30, and the predicted available resource in the subsequent period is in real-time idle. In the case of prediction of the period of available resources, a confidence such as 0.98 can also be generated for the predicted available resource. The near view layer is the final execution layer of the scheduling result, and the data accuracy requirement is the highest.

[0063] Mid view layer C2: can manage the predictive state within 30-120 minutes. As shown in the figure, the soft appointment in the dashed box and the predicted available window. The mid view layer is the main decision space for opportunity cost calculation and forward planning. Figure 2

[0064] Far view layer C3: can manage the planning state within 2-8 hours or more. As shown in the figure, the diagonal striped planned maintenance has a planned maintenance behavior at T3 period. The far view layer is the highest priority hard constraint for excluding unavailable resources from the decision source.

[0065] It can be understood that the step of managing the spatiotemporal map includes: dividing the time axis of each charging resource into different period layers, the period layers including a near view layer, a mid view layer, and a far view layer, the near view layer being used to mark a lock of an appointment charging resource within a first period with a hard appointment; the mid view layer being used to issue a predicted available window within a second period to determine a target charging resource; the far view layer being used to mark a maintenance or failure plan within a third period as a prerequisite exclusion condition for decision; wherein the length of the third period is greater than the length of the second period; the length of the second period is greater than the length of the first period.

[0066] Before the opportunity cost calculation is performed, the system first performs a preliminary screening on the mid view layer, and filters using the information of the far view layer to form an effective candidate set.

[0067] The prediction confidence is a quantitative index for evaluating the possibility of a predicted available window actually occurring in the future. The prediction confidence can be statistically and dynamically adjusted based on the performance of the charging resource or the target mobile device in the same period in history.

[0068] ​​The prediction confidence can be used as a key correction factor in the calculation of the opportunity cost index. A window with a prediction but low confidence has a high actual risk, so it will be penalized in the calculation of its opportunity cost index, making its total cost higher and thus being ranked behind windows with high confidence in the decision-making process. This makes the decision-making process not only consider the optimal path, but also the reliability of the path.

[0069] To ensure the accuracy of the decision-making and the robustness of the system in the dynamic environment of concurrent scheduling of multiple mobile devices, the application designs a complete reservation and state management mechanism.

[0070] A soft reservation is a rectangle on the C2 resource in the figure. It is a temporary and time-limited label. When the system calculates the opportunity cost index of multiple candidate windows for a device, soft reservations are applied to these windows. This is equivalent to a temporary intention lock, which can prevent multiple scheduling processes from simultaneously considering the same window as the optimal solution and causing conflicts.

[0071] A hard reservation is a rectangle on the C1 resource in the figure. When the opportunity cost index calculation is completed and the final target is selected, the soft reservation is upgraded to a hard reservation. It is a formal and binding resource lock that is recorded in the near view layer and notifies all other devices that the period is unavailable. This soft-to-hard conversion process ensures the independence, accuracy of the decision-making process, and uniqueness of the results.

[0072] In some embodiments, the smart park mobile device resource dynamic scheduling method further comprises: in the case that a mobile device completes charging in advance and releases the charging resource, clearing the hard reservation label of the released charging resource in the near view layer, and pushing the updated idle period of the released charging resource to a device queue that is waiting for charging or whose opportunity cost index calculation result does not meet the preset condition, as a potential immediately executable low-cost charging opportunity in the device queue.

[0073] When a device completes charging in advance, the system immediately releases the remaining fragmented time back to the resource pool and can actively push it to other devices with urgent or temporary needs, realizing the extraction of resource value.

[0074] When a high-level plan changes, the change must be automatically and cascadingly transmitted to the low level. For example, when an administrator inputs a maintenance plan at the C3 layer, the system must automatically delete the predicted window that conflicts with it in the C2 layer and forcibly revoke the hard reservation that has been generated in the C1 layer. This mechanism guarantees the strong consistency and real-time accuracy of the space-time map data, so that any scheduling decision is made based on the latest and most accurate global information, greatly improving the robustness of the entire system.

[0075] In some embodiments, a soft reservation with timeliness is imposed on the candidate window being evaluated in the process of calculating and comparing the opportunity cost index; after the target charging resource is determined, the soft reservation is upgraded to a hard reservation in the near view layer, and the soft reservations of other charging resources not selected are released.

[0076] When a mobile device ends charging in advance, the system immediately clears the hard reservations in the near view layer (C1) and pushes the fragmented idle period just released to the devices in the queue or not meeting the conditions as a low-cost charging opportunity that can be immediately preempted, realizing resource second-level reuse.

[0077] To prevent scheduling conflicts, any new maintenance plan added in the far view layer (C3) will automatically cascade to delete the conflict windows in the middle view layer (C2) and cancel the hard reservations already locked in the near view layer (C1), ensuring global consistency of the space-time map; at the same time, the system imposes a soft reservation with timeliness on all candidate windows when evaluating the opportunity cost index, and once the target is selected, it is upgraded to a hard reservation and the remaining soft reservations are cancelled, avoiding race conditions and ensuring real-time and accurate decision-making.

[0078] The present application not only focuses on the optimality of single scheduling, but also pursues the adaptability and global optimization of the entire system. The size of the preset power threshold is inversely proportional to the number of charging resources corresponding to the predicted available windows in the middle view layer. The triggering time of charging can be linked to the tightness of the overall field resources, forming a macro negative feedback regulation loop: when resources are tight, devices are prompted to charge as soon as possible; when resources are abundant, devices are allowed to use more power to perform tasks.

[0079] Finally, the system selects a device that minimizes the sum of the current task cost and the future charging opportunity cost. This achieves local optimization from the charging link to the task-charging whole link, which is a key step to maximize global efficiency.

[0080] The steps of the method of the present application will be described in detail below in conjunction with actual cases.

[0081] For example, a large intelligent park has 3 charging stations (C1, C2, C3), each with multiple charging piles. The park has AGV performing a warehouse-to-workstation (G2R) transfer task. The main task areas include: the warehouse-in area (A area), the warehouse-out area (B area), and the high-density storage area (H area). The current time is 9:00 am.

[0082] Step 1: Construction and management of a three-level space-time map.

[0083] The system continuously manages a dynamic space-time map of all charging resources (N1, N2, N3).

[0084] The administrator inputs in the background: charging station N3 is unavailable due to line failure on Friday afternoon (14:00-16:00) this week. After receiving the instruction, the system will mark a data maintenance on the C3 layer of all charging piles in N3. This mark is the highest priority, and any subsequent scheduling decisions must unconditionally meet the content of this maintenance mark. The data structure of this layer can be simple (resource ID, start time, end time, event type).

[0085] The data analysis module of the system will release a predicted available window on the C2 layer of N1-P1 from 9:30 to 10:30 based on historical data, such as the efficiency of charging pile N1-P1 during weekdays from 9:30 to 10:30 in the past month, which is 95% of the time.

[0086] To determine the reliability of this prediction, each predicted available window is assigned a prediction confidence. This confidence is a floating-point number between 0 and 1, such as 0.95. If the system predicts that N1-P1 is available at this time one day, but is actually temporarily occupied, the confidence will be slightly increased in the future. This confidence is a key correction factor in opportunity cost calculation.

[0087] In the near view layer, the system has allocated AGV-007 the integral of N2-P2 charging pile from 9:15 to 9:45. Then on the C1 layer, the hard reservation will be marked as hard reservation: AGV-007. This is a lock that cannot be reserved by all other devices. Through Internet of Things communication with the charging pile, the system knows in real time that N1-P2 is currently in an idle state.

[0088] Step 2: Scheduling trigger.

[0089] The system continues to monitor the status of AGV-005. Currently, AGV-005 has 45% battery power and is taking goods from area A to area B. The system predicts that after AGV-005 arrives at area B to unload, its next most likely task is to take an empty shelf from the storage area next to area B and return to area A. According to the data analysis module's path length, loading weight, etc., it is estimated that completing this B-A task will consume about 15% of the warehouse.

[0090] The system also scans the C2 layer of all charging resources in the entire field and finds that most predicted available windows have high confidence and low subsequent opportunity cost, indicating that charging resources are abundant. In this case, the system dynamically adjusts the threshold to a lower 25%. Conversely, if the system finds that charging resources are tight in the next hour (for example, multiple charging piles have maintenance plans), it will increase the threshold to 35% to encourage AGVs to plan charging earlier and avoid the dilemma of parallel no charging or high-cost charging.

[0091] Since 30%>25%, the schedule does not trigger. AGV-005 continues to execute the B-A task. After AGV-005 finishes the task and returns to A area, the current power is 30% after iteration. Its next task is to pick up goods from A area to H area, which is expected to consume 12% power. At this time, it is judged that 30%-12%=18%. Since 18%<25%, the shortcut scheduling trigger module is activated, and a charging scheduling process is initiated for AGV-005.

[0092] Step 3: Opportunity cost optimal decision.

[0093] The system receives a request for scheduling AGV-005. At this time, AGV-005 is located in A area. First, the space-time map can be queried. It scans all the medium layers of the charging piles to find predicted available windows that start within 30-120 minutes in the future. It automatically fixes those windows that conflict with the long-term maintenance plan. Suppose it finds two main candidates.

[0094] Candidate 1: Charging pile N1-P1, located near A area. C2 layer shows a predicted available window starting at 9:40, with sufficient duration and a prediction confidence of 0.98.

[0095] Candidate 2: Charging pile N2-P3, located near B area. C2 layer shows a predicted available window starting at 9:50, with sufficient duration and a prediction confidence of 0.92.

[0096] According to historical data, which area has the highest task frequency and value in the time period of 10:00-11:00 am (i.e. the time period after AGV-005 completes charging). The analysis result shows that it is the outbound area B. Therefore, B area is determined as the dynamic high-frequency task area for this scheduling.

[0097] Through calculation, the opportunity cost indices of the adjusted candidate 1 and candidate 2 are 63.78 and 55.98 respectively. Therefore, the pre-guidance cost of going to N2-P3 charging is higher, but it greatly reduces the opportunity cost after charging, i.e. it can reach the next high-value task area faster. Therefore, the system selects B.

[0098] During the opportunity cost index calculation process, the system has triggered temporary, short-lived (such as 30 seconds) soft reservation markers for candidate 1 and candidate 2. This prevents another AGV from also considering the same window as the optimal solution at the moment AGV-005 makes a decision, thereby causing a conflict.

[0099] When candidate 2 is finally selected, it will immediately send a command to N2-P3 from 9:50 to 10:25, update its state on the near layer to hard reservation: AGV-005. At the same time, the system releases the soft reservation imposed on candidate 1, and its capacity is normally scheduled by other devices.

[0100] The system sets a reservation compliance checkpoint for AGV-005's hard reservation (starting at 9:50), for example at 9:45. At this moment, the system will re-predict its arrival time according to its real-time position and speed. If it is expected to arrive at 9:53 due to sudden congestion, which exceeds the tolerance, the system can be configured to automatically cancel its reservation, release the resource to other AGVs in the waiting queue, and impose a punitive opportunity cost index on the unfaithful AGV-005, affecting its next scheduling.

[0101] AGV-005 finally charges at another charging station N1-P2, with a reservation charge from 9:55 to 10:30. But because of the good battery state, it completes charging at 10:22. After AGV-005 reports that charging is complete, the system immediately clears its hard reservation on T1 layer, and marks the newly born 8-minute fragmented time of N1-P2 from 10:22 to 10:30 as real-time waiting idle. The system can query a charging waiting queue, and if it finds an AGV that cannot find a long pulse signal charging resource, it can indicate this charging resource to it and ask if it needs a quick power-up, thereby maximizing resource efficiency.

[0102] The smart park mobile device resource dynamic scheduling method of the present application also extends the cost idea to the charging scheduling before the task allocation stage. The smart park mobile device resource dynamic scheduling method is also applied to the allocation decision of new tasks, including: when allocating tasks, determining the estimated opportunity cost index that each candidate mobile device can obtain after executing a new task based on the space-time map; and selecting the candidate mobile device with the lowest estimated opportunity cost index as the final execution mobile device for the new task.

[0103] The system issues a new task to supply from H area to A area. At this time, AGV-008 and AGV-009 are both in idle state.

[0104] The traditional decision may randomly assign or assign AGV-008 with a smaller ID. The present decision is that the task management system will initiate a pre-play cost request to the system.

[0105] The final system considers that the task is assigned to the AGV-008, not only the current task is completed, but also a more favorable position is initiated in the subsequent charging period, that is, the overall and queue of the task cost and the future charging opportunity cost. Therefore, the task is assigned to the AGV-008. This realizes the promotion of efficiency planning from single-point optimization to global optimization.

[0106] In summary, the application solves the charging problem of intelligent park mobile devices in a complex dynamic environment by constructing a three-layer space-time map containing a long-range layer, a medium-range layer and a short-range layer, and on this basis, taking the computer cost index as the core, combining dynamic high-frequency task area prediction, prediction confidence modification, software and hardware reserve management, cross-state consistency and hierarchical linkage of guarantee and task allocation, so that the scheduling decision of the charging problem of intelligent park mobile devices in a complex dynamic environment is no longer isolated, but serves the long goal of maximizing the efficiency of the whole life cycle operation of the device, and realizes the real global optimization scheduling.

[0107] The intelligent park mobile device resource dynamic scheduling system provided by the application is described below, and the intelligent park mobile device resource dynamic scheduling system described below can be correspondingly referred to the intelligent park mobile device resource dynamic scheduling method described above.

[0108] As shown in FIG. 3, the intelligent park mobile device resource dynamic scheduling system of the embodiment of the application includes a management module 310, a first processing module 320 and a second processing module 330.

[0109] The management module 310 is used for managing a space-time map recording a plurality of charging resources, and the space-time map includes a plurality of available time periods of each charging resource on a future time axis;

[0110] The first processing module 320 is used for determining a next task to be executed by a target mobile device when the target mobile device is in a schedulable state, and estimating energy consumption required for completing the next task;

[0111] The second processing module 330 is used for selecting a target charging resource for the target mobile device based on the space-time map when a result of subtracting the estimated energy consumption from a current power of the target mobile device is less than a preset power threshold; and the target mobile device travels from the target charging resource to a high-frequency task area corresponding to a target time period after the target mobile device completes charging at the target charging resource, and an opportunity cost index required for the target mobile device is the lowest.

[0112] The intelligent park mobile device resource dynamic scheduling system provided by the embodiment of the present application solves the charging problem of the intelligent park mobile device in a complex dynamic environment by constructing a multi-layer space-time map, taking a computer cost index as the core, combining dynamic high-frequency task area prediction and hierarchical linkage of task allocation, so that each charging decision is no longer isolated, but serves the long goal of maximizing the whole life cycle operation efficiency of the device, and realizes real global optimization scheduling.

[0113] Figure 4 An example of an entity structure diagram of an electronic device is shown in Figure 4 As shown, the electronic device can include a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 complete mutual communication through the communication bus 440. The processor 410 can call the logical instructions in the memory 430 to execute the intelligent park mobile device resource dynamic scheduling method, which includes: managing a space-time map recording a plurality of charging resources, the space-time map including a plurality of available time periods of each charging resource on a future time axis; when the target mobile device is in a schedulable state, determining the next task to be executed by the target mobile device and estimating the energy consumption required to complete the next task; in the case that the result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold, selecting a target charging resource for the target mobile device based on the space-time map; after the target mobile device completes charging at the target charging resource in the target time period, the target mobile device travels from the target charging resource to the high-frequency task area corresponding to the target time period with the lowest opportunity cost index.

[0114] In addition, the logical instructions in the memory 430 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0115] In another aspect, the present application also provides a computer program product comprising a computer program, which can be stored on a non-transitory computer readable storage medium, and the computer program is executable by a processor to enable a computer to perform the method of dynamically scheduling mobile device resources in a smart park, which comprises: managing a time-space map recording a plurality of charging resources, the time-space map comprising a plurality of available time periods of each charging resource on a future time axis; when a target mobile device is in a schedulable state, determining a next task to be performed by the target mobile device and estimating the energy consumption required to complete the next task; in the case that the result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold, selecting a target charging resource for the target mobile device based on the time-space map; after the target mobile device completes charging at the target charging resource in a target time period, the target mobile device has the lowest opportunity cost index required to travel from the target charging resource to a high-frequency task area corresponding to the target time period.

[0116] In another aspect, the present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, and the computer program is executable by a processor to implement the method of dynamically scheduling mobile device resources in a smart park, which comprises: managing a time-space map recording a plurality of charging resources, the time-space map comprising a plurality of available time periods of each charging resource on a future time axis; when a target mobile device is in a schedulable state, determining a next task to be performed by the target mobile device and estimating the energy consumption required to complete the next task; in the case that the result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold, selecting a target charging resource for the target mobile device based on the time-space map; after the target mobile device completes charging at the target charging resource in a target time period, the target mobile device has the lowest opportunity cost index required to travel from the target charging resource to a high-frequency task area corresponding to the target time period.

[0117] The apparatus embodiments described above are merely illustrative, wherein the units shown as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0118] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0119] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for dynamic resource scheduling of mobile equipment in a smart park, characterized in that: The method is applied to a management scheduler, and the method includes: Managing a spatiotemporal map recording a plurality of charging resources, wherein the spatiotemporal map includes a plurality of available time periods for each charging resource on a future time axis; When the target mobile device is in a schedulable state, determining a next task to be performed by the target mobile device and estimating energy consumption required to complete the next task; When the result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold, selecting a target charging resource for the target mobile device based on the spatiotemporal map; after the target mobile device completes charging at the target charging resource during the target time period, the opportunity cost index required for the target mobile device to travel from the target charging resource to the high-frequency task area corresponding to the target time period is the lowest; The steps of managing the spatiotemporal map include: The time axis of each charging resource is divided into different time period layers, including a near-term layer, a mid-term layer, and a long-term layer. The near-term layer is used to lock the reserved charging resources in the first time period with a hard reservation mark; the mid-term layer is used to publish the predicted available window in the second time period to determine the target charging resource; and the long-term layer is used to mark the maintenance or fault plan in the third time period as a prerequisite for decision-making. Among them, the duration of the third time period is greater than the duration of the second time period; the duration of the second time period is greater than the duration of the first time period; the high-frequency task area is a dynamic geographic fence, and the location and range of the high-frequency task area are predicted and updated based on the historical task heat map within the time span of the mid-view layer; the size of the preset power threshold is inversely proportional to the number of charging resources corresponding to the predicted available window in the mid-view layer.

2. The method according to claim 1, characterized in that The method is also applied to the allocation decision of new tasks, including: During task allocation, determining, based on the spatiotemporal map, an estimated opportunity cost index that each candidate mobile device can obtain after completing a new task; The candidate mobile device with the lowest estimated opportunity cost index is selected as the final execution mobile device for the new task.

3. The method according to claim 2, characterized in that The selecting a target charging resource for the target mobile device based on the spatiotemporal map includes: Screening out all predicted available windows in the mid-view layer that are not marked as unavailable by the distant view layer and have sufficient duration as candidate windows in a candidate set; When evaluating each candidate window in the candidate set, calculating the opportunity cost index of each candidate window; The charging resource corresponding to the candidate window with the lowest opportunity cost index is determined as the target charging resource and a target time period is determined.

4. The method according to claim 3, characterized in that The opportunity cost index is determined based on at least one of the following: the predicted travel time for the target mobile device to reach the charging resource corresponding to the candidate window; the predicted duration to complete charging in the candidate window; and the predicted travel time from the charging resource to the high-frequency task area after charging is completed.

5. The method according to claim 3, characterized in that The method further includes: applying a time-sensitive soft reservation to the candidate window being evaluated during the process of calculating and comparing the opportunity cost index; after determining the target charging resource, upgrading the soft reservation to a hard reservation located in the near-view layer, and releasing the soft reservations of other unselected charging resources.

6. The method according to claim 5, characterized in that The method further includes: when a mobile device completes charging in advance and releases charging resources, clearing the hard reservation mark of the released charging resources in the near-view layer, and pushing the updated idle time period of the released charging resources to a device queue that is waiting for charging or whose opportunity cost index calculation result does not meet preset conditions, as a potential low-cost charging opportunity that can be immediately executed in the device queue.

7. A dynamic scheduling system for mobile device resources in a smart park, characterized in that: include: a management module, configured to manage a spatiotemporal map recording a plurality of charging resources, wherein the spatiotemporal map includes a plurality of available time periods for each charging resource on a future time axis; The management module is further configured to divide the time axis of each charging resource into different time period layers, the time period layers including a near-term layer, a mid-term layer, and a long-term layer. The near-term layer is configured to lock a reserved charging resource within a first time period with a hard reservation mark; the mid-term layer is configured to publish a predicted available window within a second time period to determine the target charging resource; and the long-term layer is configured to mark a maintenance or fault plan within a third time period as a prerequisite for decision-making. The duration of the third time period is greater than that of the second time period; and the duration of the second time period is greater than that of the first time period. A first processing module is configured to determine a next task to be performed by the target mobile device when the target mobile device is in a schedulable state, and estimate energy consumption required to complete the next task; a second processing module configured to select a target charging resource for the target mobile device based on the spatiotemporal map when a result of subtracting the estimated energy consumption from the current power of the target mobile device is less than a preset power threshold; and wherein after the target mobile device completes charging at the target charging resource during a target period, the opportunity cost index required for the target mobile device to travel from the target charging resource to a high-frequency task area corresponding to the target period is the lowest; The high-frequency task area is a dynamic geographic fence, and the location and range of the high-frequency task area are predicted and updated based on the historical task heat map within the time span of the mid-view layer; the size of the preset power threshold is inversely proportional to the number of charging resources corresponding to the predicted available window in the mid-view layer.

Images