Intelligent device state prediction and resource elastic scheduling method for digital and intelligent space

Abstract

The application belongs to the technical field of Internet of Things resource scheduling, and provides a method for predicting the state of Internet of Things equipment and elastically scheduling resources in a digital and intelligent space, which comprises collecting equipment operation monitoring data and current resource residual value, and predicting state tension; basic priority score is calculated in combination with tension and residual resources; physical topology, workflow dependence and historical fault concurrent data are collected to construct a device correlation network in the form of a directed acyclic graph, according to which the network level depth of each device is determined and converted into a priority deduction coefficient; the final score is obtained by deducting the deduction coefficient from the basic score, and takes effect when the deviation threshold limit is met, otherwise the deduction is zero; finally, the devices are screened and resource scheduling is performed according to the final score sorting and in combination with the tension prediction value; the application effectively avoids resource allocation imbalance and improves the robustness and continuous service capability of the system.

Description

Technical Field

[0001] This invention belongs to the field of IoT resource scheduling technology, and particularly relates to a method for predicting the status of IoT devices and elastically scheduling resources for digital spaces. Background Technology

[0002] In cutting-edge application scenarios such as digital twins and intelligent operation and maintenance, the prediction of IoT device status and elastic scheduling of resources for digital spaces are the core support for maintaining the overall stability and operational efficiency of the system.

[0003] Currently, IoT device resource scheduling technologies generally rely on the current load threshold or basic priority rules of the devices to allocate resources. Some solutions introduce time series prediction models to predict the future state of the devices and adjust the scheduling strategy accordingly. At the same time, some technologies attempt to incorporate the physical connection topology and data workflow dependencies between devices to simulate the potential impact of fault propagation paths or data flow on resource demand.

[0004] However, these methods have inherent limitations. Their device status prediction and topology analysis are often separated into independent steps, failing to achieve the coordinated integration of multi-dimensional information such as device operational stress, current resource availability, and network hierarchical location. This results in the inability to construct a dynamic and comprehensive scheduling priority assessment system. In particular, when the predicted stress values ​​of multiple devices are close, existing technologies lack the ability to fine-tune control at the depth of network hierarchy. They cannot identify the inherent logic that deeper-level devices should have their priority appropriately reduced due to their higher risk of fault propagation, leading to an imbalance in resource allocation. That is, edge devices with limited impact range may consume too many resources, while core devices on the critical path cannot obtain timely support, ultimately weakening the system's robustness and continuous service capability. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for predicting the status of IoT devices and elastically scheduling resources in digital spaces, thus solving the aforementioned problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting the status of IoT devices and elastically scheduling resources in a digital space, the method specifically comprising: Collect operational monitoring data of IoT devices and current remaining resource values ​​within the digital space; Based on the monitoring data of IoT devices, the status of IoT devices is predicted, and the predicted value of the IoT device status stress is determined. The basic priority score of IoT devices is determined based on the predicted stress level of IoT devices and the current remaining resources. The physical connection topology, data workflow dependencies, and historical fault concurrency statistics of IoT devices within the digital intelligence space are collected. Based on these data, an IoT device relationship network is constructed. This network is a directed acyclic graph, with nodes connected by directed edges. Obtain the location of each node in the IoT device relationship network, and determine the IoT device network hierarchy depth based on the location of each node in the IoT device relationship network; Determine the priority reduction coefficient of IoT devices based on the network layer depth of IoT devices; The final priority score of the IoT device is determined based on the IoT device priority reduction coefficient and the IoT device basic priority score. Based on the final priority score of IoT devices and the predicted stress level of IoT devices, resources for IoT devices are scheduled.

[0007] Based on the above technical solutions, the present invention also provides the following optional technical solutions: Further technical solution: The determination of the predicted stress level of IoT devices specifically includes: Through the formula: ; Determine the predicted value of the status tension of IoT devices ; In the above formula, This refers to the predicted stress level of the i-th IoT device at time t+1 in the future. This refers to the standardized value of the IoT device operation monitoring data at time t-k+1 for the i-th IoT device. This refers to the weight coefficient corresponding to the k-th historical moment, where k refers to the number of historical moments.

[0008] Further technical solution: The determination of the basic priority score for IoT devices: Through the formula: ; Determine the basic priority score for IoT devices ; In the above formula, This refers to the basic priority score of the i-th IoT device. This refers to the predicted stress level of the i-th IoT device at time t+1 in the future. This refers to the current remaining resource value of the i-th IoT device. This refers to the maximum resource value of the i-th IoT device. All are weighting coefficients, and .

[0009] Further technical solution: Determining the network layer depth of IoT devices specifically includes: Through the formula: ; Determine the network layer depth of IoT devices ; In the above formula, This refers to the network layer depth of the i-th IoT device. This refers to the network layer depth of the j-th IoT device, and E refers to the set of directed edges from node j to node i. In this context, the depth of the central source node in the IoT device relationship network is defined as 0, and the central source node refers to the root node with an in-degree of 0.

[0010] Further technical solution: The determination of the priority reduction coefficient for IoT devices specifically includes: Through the formula: ; Determine the priority reduction factor for IoT devices ; In the above formula, This refers to the network layer depth of the i-th IoT device. This refers to the maximum network hierarchy depth of all nodes in the interconnected network of IoT devices. This refers to the adjustment factor that affects the outcome. This is a constant term.

[0011] Further technical solution: The determination of the final priority score for IoT devices specifically includes: Through the formula: ; Determine the final priority score for IoT devices ; In the above formula, This refers to the basic priority score of the i-th IoT device. This refers to the priority reduction coefficient of the i-th IoT device.

[0012] Further technical solution: When determining the final priority score of IoT devices, the specific limiting conditions include: ; in, This refers to the arithmetic mean of the basic priority scores of all IoT devices. This refers to the deviation threshold; If the conditions are not met when determining the final priority score of IoT devices, the priority reduction coefficient for IoT devices will be adjusted. The value is set to 0.

[0013] Further technical solution: the deviation threshold The specific methods of obtaining it include: Through the formula: ; Determine the availability ratio of resource scheduling ; In the above formula, This refers to the characteristic value of currently schedulable resources. This refers to the maximum value of the scheduled resources; Based on the mapping function, the resource scheduling availability ratio is calculated. Mapped to deviation threshold .

[0014] Further technical solution: The method of scheduling IoT device resources based on the final priority score and the predicted state stress of IoT devices specifically includes: Based on the final priority score of IoT devices, the IoT devices are ranked to determine the priority sequence for IoT device resource scheduling; Based on the priority sequence of IoT device resource scheduling, resource scheduling is performed on IoT devices according to the predicted value of IoT device status tension.

[0015] Further technical solution: The method of scheduling resources for IoT devices based on the priority sequence of IoT device resource scheduling and the predicted value of IoT device status tension specifically includes: Based on the predicted stress levels of IoT devices, a set of IoT devices that require resource scheduling is selected. Based on the priority sequence of IoT device resource scheduling, resource scheduling is performed on IoT devices in the set of IoT devices that need resource scheduling.

[0016] This invention provides a method for predicting the status of IoT devices and elastically scheduling resources in digital spaces, which has the following advantages compared with existing technologies: This invention dynamically adjusts device priorities by integrating device status prediction, remaining resource quantity, and network layer depth information, thereby achieving elastic resource scheduling. It realizes the collaborative integration of multi-dimensional information such as device operating stress, current remaining resource quantity, and network layer position, and constructs a dynamic and comprehensive scheduling priority evaluation system, which effectively avoids resource allocation imbalance and improves the robustness and continuous service capability of the system. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the method for predicting the status of IoT devices and elastically scheduling resources in a digital space, as provided by the present invention.

[0018] Figure 2 This is a schematic diagram of the S80 process provided by the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0021] Please see Figure 1 The present invention provides a method for predicting the state of IoT devices and elastically scheduling resources in a digital space, comprising the following steps: S10: Collect operational monitoring data of IoT devices and current remaining resource values ​​within the digital space; S20: Based on the operation monitoring data of IoT devices, predict the status of IoT devices and determine the predicted value of the IoT device status stress. S30: Determine the basic priority score of IoT devices based on the predicted value of IoT device status tension and the current remaining value of resources; S40: Collect the physical connection topology, data workflow dependencies, and historical fault concurrency statistics between IoT devices in the digital intelligence space. Based on these data, construct an IoT device relationship network. This network is a directed acyclic graph, with nodes connected by directed edges. Furthermore, if loops exist, they are broken according to physical logic. S50: Obtain the location of each node in the IoT device relationship network, and determine the IoT device network layer depth based on the location of each node in the IoT device relationship network; S60: Determine the priority reduction coefficient of IoT devices based on the network layer depth of IoT devices; S70: Determine the final priority score of the IoT device based on the IoT device priority reduction coefficient and the IoT device basic priority score; S80: Based on the final priority score of IoT devices and the predicted state stress of IoT devices, the resources of IoT devices are scheduled; Among them, the digital space refers to a complex system environment that integrates the physical world and the digital world and is driven by intelligent technology. In this environment, various IoT devices are widely deployed and generate massive amounts of data. At the same time, through intelligent analysis and decision-making, the perception, control and optimization of physical entities are realized.

[0022] IoT devices refer to various physical devices deployed in digital spaces that have sensing, computing, communication, and execution capabilities. These devices can be sensors, actuators, smart terminals, edge computing devices, etc. They are interconnected through networks and together constitute the infrastructure of digital spaces.

[0023] IoT device operation monitoring data refers to data collected in real time or periodically through sensors, log records or other monitoring methods, which reflects the current operating status and performance of IoT devices. This data can include CPU utilization, memory usage, network bandwidth, temperature, power consumption, error rate, etc.

[0024] The current remaining resource value refers to the amount of resources currently available to IoT devices or the systems they depend on. These resources can be computing resources (such as the number of CPU cores and memory size), storage resources, network bandwidth, power, etc., and their remaining value reflects the carrying capacity of the device or system.

[0025] Physical connection topology refers to the actual physical connection structure between IoT devices in the digital space, such as the connection relationship formed by network cables, wireless signals, etc. This topology describes the direct communication path between devices.

[0026] Data workflow dependencies refer to the dependencies between IoT devices in data processing and business logic; for example, if the data processing result of device A is the input for device B to perform the next operation, then device B depends on device A.

[0027] Historical failure concurrency statistics refer to statistical information such as the frequency and type of failures of IoT devices in the digital space and the probability of different devices failing at the same time over a period of time. This data helps to assess the potential risk correlation between devices.

[0028] The network hierarchy depth of IoT devices refers to the hierarchical position of IoT devices in the network of relationships between IoT devices. Generally, the farther a device is from the central source node of the network, the greater its hierarchy depth, which can reflect the importance of the device in the whole system or the potential impact range of its failure propagation.

[0029] Specifically, firstly, in step S10, it is necessary to collect operational monitoring data and current remaining resource values ​​of IoT devices within the intelligent space. This data collection can be achieved in several ways. For example, a human operator can periodically log into the management interface of each IoT device to manually record operational monitoring data such as CPU utilization, memory usage, and network bandwidth, as well as the currently available resource quantity. Alternatively, a lightweight agent program can be deployed on each IoT device, which periodically collects the device's operational status indicators and resource usage, and reports this data to the central data collection server via the network. Furthermore, network sniffing or traffic analysis tools can be used to intercept and parse communication data between IoT devices at the network layer, extracting operational monitoring data and resource usage information.

[0030] Secondly, in step S20, the status of the IoT devices is predicted based on the collected IoT device operation monitoring data to determine the predicted stress level of the IoT devices. This prediction process can employ various technical means. For example, a series of static thresholds can be set; when a key operating indicator of the IoT device (such as CPU utilization) exceeds the preset threshold for a continuous period of time, its status is predicted to tend towards stress. Alternatively, simple statistical methods can be used, such as calculating the moving average of the IoT device's operating indicators over a past period and judging the future stress trend based on the deviation between the current value and the moving average.

[0031] Next, in step S30, the basic priority score of the IoT device is determined by combining the predicted state stress value and the current remaining resource value. This score can be determined by a direct weighted summation method, for example, multiplying the predicted state stress value by a weighting coefficient, multiplying the reciprocal of the current remaining resource value by another weighting coefficient, and then adding the two together to obtain the basic priority score. Alternatively, it can be implemented using a preset lookup table, which directly maps different intervals of predicted state stress values ​​and resource remaining value values ​​to a preset basic priority score.

[0032] Subsequently, in step S40, it is necessary to collect the physical connection topology, data workflow dependencies, and historical fault concurrency statistics among IoT devices within the digital intelligence space, and construct a network of relationships between IoT devices based on this data. This network is constructed as a directed acyclic graph, with nodes connected by directed edges. The construction process may include: first, obtaining physical connection information between devices through manual configuration or network device discovery protocols (such as LLDP, CDP); second, identifying data workflow dependencies between devices by analyzing application logs, business process definitions, or manual configuration; and finally, extracting fault concurrency statistics from historical fault management systems or operation and maintenance records. After integrating this information, a graph database or in-memory graph structure can be used to represent this network of relationships, and necessary loop detection and violation handling can be performed to ensure its directed acyclic properties.

[0033] Further, in step S50, the location of each node in the IoT device inter-device relationship network is obtained, and the IoT device network hierarchy depth is determined based on the location. This hierarchy depth can be determined starting from one or more preset "core" nodes in the network. Using breadth-first search (BFS) or depth-first search (DFS) algorithms, the shortest path length from other nodes to these core nodes is calculated, and this length is defined as the hierarchy depth. For example, a node with an in-degree of zero can be defined as the initial hierarchy (depth 0), and then the hierarchy depth of its directly downstream nodes is calculated layer by layer.

[0034] Based on this, in step S60, the priority reduction coefficient of the IoT devices is determined according to the network layer depth. This coefficient can be determined using a simple linear relationship, such as multiplying the layer depth by a fixed scaling factor. Alternatively, it can be mapped using a preset function curve, so that the greater the layer depth, the greater the reduction coefficient.

[0035] Then, in step S70, the final priority score of the IoT device is determined based on the IoT device priority reduction coefficient and the IoT device basic priority score. This final score can be determined by direct subtraction, that is, subtracting the priority reduction coefficient from the IoT device's basic priority score. For example, if the basic priority score is 80 and the reduction coefficient is 10, then the final priority score is 70.

[0036] Finally, in step S80, resources for the IoT devices are scheduled based on their final priority scores and predicted state stress levels. This scheduling process may include: first, sorting all IoT devices from highest to lowest according to their final priority scores; then, allocating the required resources to each device sequentially from the sorted list until available resources are exhausted or the needs of all devices are met. When allocating resources, devices with both high final priority scores and high predicted state stress levels may be given priority.

[0037] This invention integrates IoT device status prediction with resource remaining quantity calculation as a fundamental priority, and quantifies the device network hierarchy depth based on a directed acyclic graph structure. It introduces a penalty coefficient to correct the priority of deeper devices, ensuring that resource scheduling priorities reflect both the urgency of the devices themselves and the breadth of their impact in the fault propagation chain. Simultaneously, a dynamic deviation threshold defines the competition range, applying hierarchical penalty only between devices with similar scores, avoiding excessive punishment and ensuring the rationality and stability of scheduling. Finally, it combines predicted values ​​to select devices for scheduling, achieving flexible resource allocation towards high-urgency, high-impact devices, effectively improving the overall reliability and resource utilization efficiency of the digital space.

[0038] Preferably, the present invention further proposes the method of determining the predicted value of the state tension of IoT devices, specifically including: Through the formula: ; Determine the predicted stress level of IoT devices; In the above formula, This refers to the predicted stress level of the i-th IoT device at time t+1 in the future. This refers to the standardized value of the IoT device operation monitoring data at time t-k+1 for the i-th IoT device. The term refers to the weight coefficient corresponding to the k-th historical moment, where k refers to the number of historical moments. Among them, determining the predicted stress level of IoT devices aims to quantify the degree of resource pressure or operational risk that IoT devices may face at a certain point in the future. Its role is to provide a key input parameter for subsequent priority scoring and resource scheduling, enabling the system to more accurately assess the urgency of the devices.

[0039] The implementation methods may include, but are not limited to: predicting equipment operation data based on time series analysis models (such as ARIMA, LSTM, etc.), or estimating by combining historical data trends with statistical methods (such as moving average, exponential smoothing, etc.).

[0040] The above formula defines a method for calculating the predicted stress level of IoT devices by weighted summation of historical standardized data. It comprehensively considers data from multiple historical moments and assigns different weights to different historical moments to more fully and precisely reflect the evolution trend and stress level of device status.

[0041] IoT device status stress prediction value It is one of the core outputs of this method, used to characterize the potential stress level of IoT devices in the future. The higher the value, the more stressed the device is, and the higher the demand for resources or the greater the risk.

[0042] The standardized value of the IoT device's operational monitoring data at time t-k+1 is used to eliminate differences in the dimensions and numerical ranges of different monitoring data, making them comparable and preventing certain indicators with larger values ​​from dominating the calculation. Common standardization methods include Max-Min Normalization, which linearly transforms the data to the interval [0,1]; or Z-score standardization, which converts the data into a distribution with a mean of 0 and a standard deviation of 1.

[0043] The weighting coefficient corresponding to the k-th historical moment is used to reflect the degree of influence of data from different historical moments on the predicted value of future state tension. Generally, historical data closer to the current moment may have a greater impact on the future state, and therefore can be assigned a higher weight; conversely, historical data from further back may have a lower weight. These weighting coefficients can be set empirically, adjusted by expert knowledge, or learned from historical data through machine learning algorithms (such as regression analysis).

[0044] Specifically, after collecting IoT device operation monitoring data, the raw data is first standardized to eliminate dimensional differences between different data sources or indicators, ensuring that all data are compared and calculated on a uniform scale. Then, using the standardized historical operation monitoring data and preset weighting coefficients, a weighted summation method is used to calculate the predicted state stress of IoT devices at future times. The weighting coefficients allow the system to differentiate based on the importance of historical data (e.g., recent data is generally more valuable than older data), thus more accurately capturing the dynamic evolution trend of IoT device states. This prediction mechanism not only considers the historical operating conditions of the devices but also ensures data consistency through standardization and highlights key historical information through weighted summation. This allows the obtained predicted state stress values ​​of IoT devices to more accurately reflect the potential resource pressure or operational risks that devices may face in the future, providing a more reliable and refined input for subsequent IoT device basic priority scoring, thereby improving the accuracy and effectiveness of the entire resource scheduling method.

[0045] Through the above technical solutions, this application can more comprehensively and precisely assess the state stress of IoT devices in the future, accurately reflect the evolution trend of device state, provide a more reliable basis for subsequent basic priority scoring of IoT devices, avoid resource scheduling deviations caused by inaccurate predictions, and enable resource scheduling strategies to respond to the actual needs of IoT devices more timely and effectively, thereby improving the intelligence and precision of IoT device resource scheduling in the digital space.

[0046] Preferably, the present invention further proposes the following method for determining the basic priority score of IoT devices: Through the formula: ; Determine the basic priority score for IoT devices ; In the above formula, This refers to the basic priority score of the i-th IoT device. This refers to the predicted stress level of the i-th IoT device at time t+1 in the future. This refers to the current remaining resource value of the i-th IoT device. This refers to the maximum resource value of the i-th IoT device. All are weighting coefficients, and ; The IoT device basic priority score aims to quantify the initial urgency or importance of a single IoT device for resource scheduling, regardless of its location in the network topology. It can serve as the initial basis for subsequent refined scheduling decisions.

[0047] The predicted stress level of IoT devices reflects the operational pressure or potential failure risks that IoT devices may face at a future point in time. Its role is to provide forward-looking information, enabling the system to intervene before problems occur. This predicted value can be calculated based on historical operational data using time series analysis models (such as ARIMA, LSTM, etc.) or machine learning models (such as support vector machines, random forests, etc.).

[0048] The current remaining resource value for an IoT device represents the amount of a specific type of resource currently available to the device. For example, for computing resources, it could be the remaining percentage of CPU utilization; for storage resources, it could be the amount of available storage space. This value is typically obtained through a real-time monitoring system, such as from a device agent via the SNMP protocol, or through telemetry data reported by the device itself.

[0049] The maximum resource limit for IoT devices represents the total amount or upper limit of a specific type of resource that an IoT device can possess. For example, the number of CPU cores, total memory size, or maximum bandwidth of a network interface for a server. This value is typically configured during device deployment or obtained from the device's specifications, and can also be automatically discovered and recorded by the system.

[0050] Weighting coefficients are used to adjust the relative importance of IoT device status stress predictions and resource availability in the basic priority score. By adjusting these coefficients, the system can be made to focus more on preventative maintenance or responsive scheduling. These coefficients can be preset based on expert experience or dynamically optimized and learned through machine learning methods (such as reinforcement learning, genetic algorithms, etc.) based on historical scheduling results.

[0051] Specifically, the first part of the above formula The predicted stress level of IoT devices was directly adopted. This predicted value was derived using the methods described above (such as weighted summation of historical operational monitoring data) and reflects the potential operational risks the devices may face in the future. A higher predicted value indicates a more stressed device status and a more urgent need for resource allocation. The second part of the formula... The system focuses on the current resource availability of IoT devices, calculating the remaining resource ratio and translating it into a contribution to the priority score. When an IoT device has a low remaining resource value, it indicates that its resources are nearing depletion, and the need for resource scheduling is more urgent. By adjusting the weighting coefficients, the system can flexibly adjust the relative importance of state tension and resource availability in the basic priority score based on actual business needs or strategy preferences. This combined approach ensures that the basic priority score for IoT devices is not only forward-looking (based on predicted values) but also timely (based on current resource availability), thus providing a more comprehensive and accurate initial assessment for subsequent refined scheduling.

[0052] Through the above technical solution, this application overcomes the limitations of prioritizing IoT devices solely based on predicted IoT device status stress values. By weightedly fusing the predicted IoT device status stress values ​​with the current remaining resource values ​​of the IoT devices, the determined basic priority score of the IoT devices can more comprehensively and accurately reflect the true urgency of resource scheduling for the IoT devices. This enables the system to not only proactively address potential device status deterioration during resource scheduling but also promptly respond to devices currently experiencing resource shortages. This avoids resource mismatch or untimely scheduling issues caused by single-dimensional assessment, significantly improving the rationality and efficiency of IoT device resource scheduling within the digital space.

[0053] Preferably, the present invention further proposes the method of determining the network layer depth of IoT devices, specifically including: Through the formula: ; Determine the network layer depth of IoT devices ; In the above formula, This refers to the network layer depth of the i-th IoT device. This refers to the network layer depth of the j-th IoT device, and E refers to the set of directed edges from node j to node i. Among them, the depth of the central source node in the IoT device relationship network is defined as 0, and the central source node refers to the root node with an in-degree of 0. The network layer depth of IoT devices is a technical characteristic used to quantify the hierarchical position of IoT devices within the network of interconnected IoT devices. It reflects the "distance" or "influence" of a device in the data flow or physical connection chain. Generally, a greater depth indicates that the device is located further downstream or at the end of the chain, and may depend on more upstream devices. This depth can be calculated using graph traversal algorithms, such as Breadth-First Search (BFS) or Depth-First Search (DFS), starting from the root node (i.e., the central source node) and exploring downwards layer by layer, marking the depth of each node. Alternatively, it can be calculated using topology sorting methods, which, during the topology sorting of a directed acyclic graph, calculate the depth of the current node based on the change in in-degree of the node and the depth information of its predecessor nodes.

[0054] The above formula provides a recursive and systematic method to calculate the network hierarchy depth of each IoT device. Its purpose is to ensure that the depth of each node is determined based on the maximum depth of all its direct predecessor nodes, thus reflecting the "deepest" dependency path of that node in the entire network and guaranteeing the accuracy and consistency of the hierarchy depth calculation. This formula can be used iteratively, starting from the central source node with a depth of 0, calculating the depth of its direct child nodes layer by layer until the depth of all reachable nodes has been calculated. Alternatively, a dynamic programming approach can be used to fill in the depth values ​​from bottom to top or top to bottom, where the depth calculation of each node depends on the calculated depth of its predecessor nodes.

[0055] E refers to the set of directed edges pointing from node j to node i. This technical feature clarifies that when calculating the depth of an IoT device network hierarchy, the directed connections in the network of relationships between IoT devices need to be considered. This emphasizes the directionality of data flow or physical dependencies, ensuring that the calculation of hierarchy depth is based on actual dependencies. When constructing the network of relationships between IoT devices, data structures such as adjacency matrices or adjacency lists can be used to store the directed edge information between nodes, facilitating a quick lookup of all predecessor nodes of a node when calculating depth. Alternatively, the set E can be constructed by traversing all edges in the network and identifying all edges pointing to node i.

[0056] The depth of the central source node in the IoT device interconnection network is defined as 0. The central source node refers to the root node with an in-degree of 0. This technical feature provides a clear starting condition and benchmark for calculating the depth of the IoT device network hierarchy. Its role is to ensure the consistency and traceability of depth calculation, because the root node with an in-degree of 0 does not depend on any other device in the network and is the starting point of the entire data flow or physical connection; defining its depth as 0 is logical. After constructing the IoT device interconnection network, all nodes can be traversed, the in-degree of each node can be calculated, and nodes with an in-degree of 0 can be identified as central source nodes, with their depth initialized to 0. Alternatively, graph algorithms, such as the preprocessing stage of topology sorting, can be used to identify all nodes without predecessor nodes as central source nodes and set their initial depth.

[0057] The proposed solution first identifies the central source node with an in-degree of 0 in the network of relationships between IoT devices. These nodes serve as the starting point of the entire network hierarchy, and their depth is explicitly defined as 0. Subsequently, for other IoT devices i in the network, their network hierarchy depth is determined based on the maximum depth of all their direct predecessor nodes j plus 1. This means that the hierarchy depth of an IoT device reflects the longest dependency path from the central source node to that device. This recursive calculation method ensures that the position of each IoT device in the network is accurately quantified, thereby accurately reflecting its relative position and importance in the data flow or physical dependency chain within the entire digital intelligence space. This method of determining hierarchy depth provides a structured and directional basis for the subsequent calculation of IoT device priority reduction coefficients, enabling priority assessment to fully consider the scope of influence of devices in the network topology and avoiding the one-sidedness of simply judging priority based on the device's own state.

[0058] By explicitly defining the depth of the central source node and employing a recursive formula to calculate the depths of other nodes, this application ensures that the hierarchical position of each IoT device in a complex interconnected network can be accurately quantified. This allows the subsequent calculation of IoT device priority reduction coefficients to fully consider the actual impact and dependencies of devices in the network topology, thereby avoiding priority scoring biases caused by inaccurate hierarchical depth assessment. Ultimately, this contributes to achieving more rational and efficient elastic scheduling of IoT device resources, improving the overall operational stability and resource utilization of the digital space.

[0059] Preferably, the present invention further proposes the determination of the IoT device priority reduction coefficient, specifically including: Through the formula: ; Determine the priority reduction factor for IoT devices ; In the above formula, This refers to the network layer depth of the i-th IoT device. This refers to the maximum network hierarchy depth of all nodes in the interconnected network of IoT devices. This refers to the adjustment factor that affects the outcome. For constant terms; The IoT device priority reduction coefficient is a parameter used to adjust the basic priority score of IoT devices. Its function is to appropriately lower the priority of an IoT device based on its hierarchical depth within the network. This coefficient aims to reflect the varying importance of IoT devices within the network; generally, the deeper the hierarchical level, the smaller its impact on core business or critical paths, and therefore its priority should be lowered accordingly. This coefficient can be determined through a pre-defined functional relationship or a mapping relationship based on expert experience or historical data analysis.

[0060] The network layer depth of the i-th IoT device represents the hierarchical position of the device in the network of relationships between IoT devices.

[0061] The maximum network layer depth of all nodes in the IoT device relationship network is used to normalize the depth of the current device, making it between 0 and 1.

[0062] The influence adjustment factor is used to control the overall impact of network hierarchy depth on priority reduction. It can be adjusted according to actual business needs and system sensitivity. For example, it can be set to 0.1, 0.5 or 1.0.

[0063] The main purpose of the constant term is to prevent the denominator from being zero, especially when the network structure is very simple and has only one level. This constant term can be a very small positive number, such as 0.001 or 1.

[0064] Specifically, based on the determined network layer depth of IoT devices, this scheme utilizes the ratio of the IoT device network layer depth to the maximum layer depth of the entire network, weighted by an influence adjustment coefficient, and introduces a constant term to avoid a zero denominator, thus obtaining a quantified priority reduction coefficient. This coefficient accurately reflects the degree of negative impact of the relative layer position of IoT devices in the network on their priority. Devices with deeper layers have greater IoT device network layer depth, resulting in a larger calculated reduction coefficient, meaning their final priority will be reduced more significantly. Conversely, devices with shallower layers have smaller reduction coefficients, and their priority is less affected. This mechanism allows the resource scheduling system to fine-tune the basic priority of IoT devices based on their structural importance in the interconnected network, avoiding the shortcomings of simply making coarse judgments based on depth, and ensuring the rationality and effectiveness of the scheduling strategy.

[0065] Through the above technical solution, this application can accurately calculate the priority reduction coefficient of IoT devices based on their hierarchical depth in the interconnected network. This allows the system to more meticulously assess the structural importance of IoT devices within the entire digital intelligence space, avoiding unreasonable priority adjustments caused by differences in network hierarchy depth. Especially in complex IoT device networks, this solution can effectively differentiate the impact of devices at different levels on system stability and business processes, ensuring that core or upstream devices receive higher priority, while downstream or non-critical devices receive appropriate priority reductions. This helps optimize resource scheduling strategies, making resource allocation more aligned with actual business needs and network topology, thereby improving the efficiency of IoT device resource scheduling and the overall stability of the system within the digital intelligence space.

[0066] Preferably, the present invention further proposes the method of determining the final priority score of IoT devices, specifically including: Through the formula: ; Determine the final priority score for IoT devices ; In the above formula, This refers to the basic priority score of the i-th IoT device. This refers to the priority reduction coefficient of the i-th IoT device; The IoT device final priority score is a comprehensive indicator used to guide the resource scheduling of IoT devices. This score aims to comprehensively measure the importance and urgency of IoT devices in the digital space, so as to make optimal scheduling decisions under limited resources. It can be a standardized value between 0 and 1, with a higher value indicating that the IoT device should receive a higher priority in resource scheduling.

[0067] This application's solution determines the final priority score of IoT devices by combining a basic priority score with a priority reduction coefficient. The method first calculates a basic priority score reflecting the urgency of the IoT device based on its own operational monitoring data and remaining resource value. Then, considering the structural importance of IoT devices in the digital intelligence space network, a priority reduction coefficient is introduced. This coefficient is determined based on the network layer depth of the IoT device; devices with greater network layer depth generally have a smaller impact on the core functions of the entire digital intelligence space, thus their priority reduction coefficient will be correspondingly larger. By subtracting the priority reduction coefficient from the basic priority score, this solution allows for fine-tuning of the initial priority of IoT devices. This adjustment mechanism ensures that resource scheduling not only considers the current urgency of the IoT device but also its strategic position within the entire digital intelligence space network. For example, a device that is urgent but located at the network edge may have its final priority appropriately reduced; while a device with slightly lower urgency but located at the network core may maintain a relatively high final priority. This comprehensive evaluation approach enables resource allocation decisions to better balance local urgency with global importance, thereby optimizing the overall resource utilization efficiency and system stability of the digital space.

[0068] Through the above technical solution, this application, in determining the final priority score of IoT devices, not only considers the urgency reflected by the predicted state tension value and the current remaining resource value of the IoT devices themselves, but also creatively introduces an IoT device priority reduction coefficient calculated based on the network layer depth of the IoT devices. This reduction mechanism enables the final priority score of IoT devices to more comprehensively and accurately reflect the overall importance of IoT devices in the digital intelligence space. It effectively solves the problem that resource scheduling based solely on the urgency of the devices may lead to insufficient resources at critical nodes or wasted resources at non-critical nodes. By applying priority reduction to devices with deeper network layers, this solution can prioritize the allocation of limited resources to core or upstream devices that have a greater impact on the overall operation of the digital intelligence space, thereby significantly improving the strategic nature and effectiveness of resource scheduling, enhancing the overall stability and resilience of the digital intelligence space, and reducing the possibility of systemic risks caused by local failures.

[0069] Preferably, the present invention further proposes that the limiting conditions for determining the final priority score of IoT devices specifically include: ; in, This refers to the arithmetic mean of the basic priority scores of all IoT devices. This refers to the deviation threshold; This constraint is used to define which devices are in the "same score / similar score" competition range. Specifically, it is achieved by calculating the absolute difference between the basic priority score of a single IoT device and the arithmetic mean of the basic priority scores of all IoT devices, and comparing this difference with a preset deviation threshold. Its purpose is to accurately filter out devices that require further fine-tuning of priority, avoiding unnecessary intervention for devices with significant priority differences. This constraint can be based on a preset fixed threshold or dynamically adjusted according to system operating status or resource availability. If the conditions are not met when determining the final priority score of IoT devices, the priority reduction coefficient for IoT devices will be adjusted. The value is set to 0.

[0070] The basic priority score is the arithmetic mean of all IoT devices. This average serves as a benchmark for measuring the overall priority level of IoT devices and provides a reference for determining whether the basic priority score of an individual IoT device is in the "same score / similar score" competition range. It is usually calculated by summing the basic priority scores of all IoT devices and then dividing by the number of IoT devices.

[0071] The deviation threshold is used to define the acceptable range of deviation between the base priority score and the average value, thereby defining the width of the competition interval for "same score / similar score". The deviation threshold can be a pre-set fixed value, such as determined by experience or system design requirements; or it can be dynamically calculated based on the current system status, resource load or distribution of base priority scores, and can be adjusted based on statistical methods (such as a multiple of the standard deviation).

[0072] Based on the aforementioned constraints, the system can identify sets of IoT devices with similar basic priority scores. These devices may exhibit similar importance in the initial assessment, thus requiring more refined strategies to differentiate their priorities to ensure fairness and effectiveness in resource allocation.

[0073] If the final priority score of an IoT device does not meet the specified conditions, the priority reduction coefficient for that device is set to 0. This means that for IoT devices with basic priority scores significantly higher or lower than the average level, the priority reduction effect brought about by network layer depth will be ignored. This ensures that the priority reduction coefficient only applies to devices that truly require fine-grained differentiation through network topology, thereby avoiding unnecessary adjustments to devices that have already been identified as high or low priority, maintaining their original basic priority assessment results, and thus improving the accuracy of resource scheduling.

[0074] This application's solution, based on the already determined basic priority scores and priority reduction coefficients of IoT devices, introduces an intelligent conditional judgment mechanism to determine the final priority score. First, the system calculates the arithmetic mean of the basic priority scores of all IoT devices. Then, for each IoT device, the system evaluates the absolute difference between its basic priority score and this average. If this difference is less than or equal to a preset deviation threshold, the IoT device is considered to be in a competitive range of "same score / similar score". For these devices in the competitive range, their previously calculated priority reduction coefficients will be applied normally, and their basic priority scores can be finely adjusted according to their network layer depth to obtain the final priority score. Conversely, if the absolute difference between the IoT device's basic priority score and the average is greater than the deviation threshold, it indicates that the device's priority is either significantly higher or significantly lower than the average level, and its priority is relatively clear, requiring no additional reduction adjustment based on network layer depth. In this case, the IoT device priority reduction coefficient will be set to 0, meaning that the device's final priority score will directly equal its basic priority score, and the influence of network layer depth on its priority is temporarily ignored. By employing this conditional priority reduction coefficient application mechanism, the proposed solution ensures that the structural factor of network hierarchy depth only functions in scenarios requiring fine-grained differentiation, namely when the basic priority scores of IoT devices are ambiguous or competitive. This avoids unnecessary adjustments to devices whose priorities are already clearly defined, thereby improving the accuracy and rationality of the final priority score. This allows resource scheduling to more precisely respond to the actual needs of IoT devices and their structural importance in the digital intelligence space.

[0075] By introducing constraints on the final priority score of IoT devices, the proposed solution avoids applying a uniform priority reduction coefficient to all IoT devices. This constraint allows the system to intelligently identify IoT devices whose basic priority scores fall within the "same / similar score" competition range. For these devices, the priority reduction coefficient resulting from network layer depth can play a role in fine-grained differentiation, providing a more accurate basis for priority ranking when the initial priority assessment results are ambiguous. For IoT devices whose basic priority scores significantly deviate from the average, i.e., those whose priorities are relatively clear, the priority reduction coefficient is set to 0, avoiding unnecessary adjustments and ensuring the validity of their original priority assessment results. This selective application strategy makes the calculation of the final priority score more accurate and reasonable, effectively solving the problem of difficulty in effectively distinguishing device priorities when basic priority scores are similar, thereby improving the accuracy and efficiency of IoT device resource scheduling in the digital space.

[0076] Preferably, the present invention further proposes the aforementioned deviation threshold. The specific methods of obtaining it include: Through the formula: ; Determine the available ratio of resources for scheduling ; In the above formula, This refers to the characteristic value of currently schedulable resources. This refers to the maximum value of the scheduled resources; Based on the mapping function, the resource scheduling availability ratio is mapped to the deviation threshold. ; The resource availability ratio is used to quantify the current resource status available for resource scheduling in the system. The currently available resource characteristics can refer to single or combined resource quantities such as the number of CPU cores, memory capacity, network bandwidth, and storage space currently available in the system. These characteristics can be obtained through real-time monitoring of system performance indicators, such as through operating system APIs, virtualization platform interfaces, or dedicated monitoring agents.

[0077] The maximum scheduling resource value represents the maximum amount of resources that the system can provide, such as the total number of CPU cores, total memory capacity, maximum network bandwidth, or total storage space. It can be defined based on the system hardware configuration or the preset resource pool limit.

[0078] The mapping function converts the calculated resource availability ratio into a specific deviation threshold. This mapping function can be implemented in various forms. For example, it can be a predefined linear function that establishes a simple direct or inverse proportional relationship between the resource availability ratio and the deviation threshold; it can also be a piecewise function that applies different conversion rules to different intervals of the resource availability ratio; or it can be a non-linear model trained on historical data, such as a neural network or support vector machine, to capture more complex mapping relationships; furthermore, it can be implemented using a lookup table method, where a table of correspondence between the resource availability ratio and the deviation threshold is pre-built, and the corresponding deviation threshold is looked up based on the current resource availability ratio.

[0079] Specifically, the system first acquires the current schedulable resource characteristic values ​​and the maximum schedulable resource value in real time, calculating the current resource availability ratio. This ratio directly reflects the resource scarcity level of the system: a high resource availability ratio indicates relatively abundant system resources, while a low ratio indicates relatively scarce resources. Subsequently, a preset mapping function converts this dynamically changing resource availability ratio into a corresponding deviation threshold. For example, when system resources are abundant, the mapping function can generate a larger deviation threshold, meaning that more IoT devices will be considered to be in the "same / similar score" competition range, allowing the priority reduction coefficient to be adjusted for a wider range of devices. Conversely, when system resources are scarce, the mapping function can generate a smaller deviation threshold, ensuring that only devices with very similar basic priority scores are included in the competition range, thus limiting the application of the priority reduction coefficient to more critical devices. This mechanism allows the deviation threshold to adaptively adjust according to the actual resource availability within the digital space, ensuring that the calculation of the final priority score for IoT devices can respond more flexibly and accurately to changes in system resources, avoiding the scheduling rigidity problem that may result from a fixed threshold.

[0080] Through the aforementioned technical solution, the deviation threshold is no longer statically fixed but can be dynamically adjusted based on the actual available resource allocation ratio within the digital space. This dynamic adjustment mechanism makes the calculation of the final priority score for IoT devices more intelligent and adaptive, effectively addressing the challenges posed by system resource fluctuations. When system resources are abundant, the definition of "same / similar scores" can be appropriately broadened, allowing more devices to participate in priority adjustments, thereby optimizing overall resource utilization. When system resources are strained, the definition can be tightened, focusing priority adjustments on the most critical devices to ensure the stable operation of core businesses. This significantly enhances the elasticity and responsiveness of IoT device resource scheduling, enabling resource allocation strategies to better adapt to the complex and ever-changing environment of the digital space, thereby improving the efficiency and accuracy of resource scheduling.

[0081] For preferred options, please refer to [link / reference]. Figure 2 The present invention further proposes to schedule IoT device resources based on the final priority score and the predicted state stress value of IoT devices, specifically including: S81: Based on the final priority score of the IoT devices, rank the IoT devices and determine the priority sequence for IoT device resource scheduling; specifically including: Based on the predicted stress levels of IoT devices, a set of IoT devices that require resource scheduling is selected. Based on the priority sequence of IoT device resource scheduling, resource scheduling is performed on IoT devices in the set of IoT devices that need resource scheduling. S82: Based on the priority sequence of IoT device resource scheduling, resource scheduling is performed on IoT devices according to the predicted value of IoT device status tension; Arranging IoT devices refers to organizing them in a specific order within the digital space according to certain sorting rules. The purpose of this operation is to provide a clear execution order for subsequent resource scheduling. Possible implementation methods include, but are not limited to: using standard sorting algorithms such as bubble sort, quicksort, and merge sort, arranging the IoT devices in descending order using their final priority score as the primary key, with higher scores indicating higher priority; or, in the case of identical scores, combining other auxiliary indicators (such as device ID, deployment time, etc.) for secondary sorting to ensure the uniqueness and determinism of the sequence.

[0082] The IoT device resource scheduling priority sequence refers to an ordered list formed by sorting all IoT devices according to their final priority scores. This sequence clearly defines the order in which IoT devices are scheduled for resources. Possible implementation methods include, but are not limited to: sorting all IoT devices in descending order of their final priority scores, with higher scores indicating higher priority; or, in the case of identical scores, performing a secondary sorting based on other auxiliary indicators (such as device ID, deployment time, etc.) to ensure the uniqueness and determinism of the sequence.

[0083] Selecting the set of IoT devices requiring resource scheduling refers to identifying those devices from all IoT devices that currently or are predicted to require additional resource support in the future. This selection process aims to focus resource scheduling on the scope and improve scheduling efficiency. Possible implementation methods include, but are not limited to: setting a threshold for the predicted state stress of IoT devices, with devices exceeding this threshold included in the set; or, combining factors such as the current resource utilization rate and historical resource demand patterns of IoT devices, and using preset strategies or algorithms for judgment and selection.

[0084] Resource scheduling for IoT devices within a set of IoT devices requiring resource allocation refers to allocating available resources to selected IoT devices according to a predetermined scheduling strategy and priority sequence. This operation aims to meet the resource needs of IoT devices, maintain their normal operation, or optimize their performance. Possible implementation methods include, but are not limited to: using a greedy algorithm to allocate resources to devices sequentially according to priority until resources are exhausted or device needs are met; or using an optimization algorithm based on constraint satisfaction to maximize overall system performance or minimize resource waste while satisfying the resource needs of all devices.

[0085] This application's solution refines the resource scheduling process for IoT devices into two stages, achieving refined and flexible scheduling of IoT device resources within the digital space. First, the system globally ranks all IoT devices based on their final priority scores, thus determining a resource scheduling priority sequence. This sequence comprehensively reflects the importance of each IoT device in the network, its own state stress level, and resource availability, providing clear guidance for subsequent resource allocation. Second, based on this priority sequence, the system further utilizes the predicted state stress values ​​of IoT devices to filter out those that genuinely require resource scheduling. This means that even if an IoT device has a high final priority score, if its predicted state stress value indicates that it has no resource needs in the near future, it will not be scheduled for resource allocation, thus avoiding resource waste. Conversely, for IoT devices with high predicted state stress values ​​and high priority, the system will prioritize resource allocation. This scheduling mechanism, combining global priority ranking and local demand filtering, ensures that resources are prioritized for allocation to the most needed and important IoT devices, while avoiding unnecessary resource allocation, thereby improving resource utilization and scheduling efficiency.

[0086] Through the above technical solution, this application effectively addresses the problem of how to rationally combine the final priority score and the predicted state stress value of IoT devices during resource scheduling. By first determining the global scheduling order based on the final priority score, and then precisely filtering based on the predicted state stress value, it ensures that resources are prioritized for allocation to IoT devices that are both important and have actual resource needs. This avoids unnecessary resource allocation to high-priority devices with no actual need, thereby significantly improving resource utilization efficiency, reducing scheduling conflicts and resource waste, and making IoT device resource scheduling in the digital space more accurate, efficient, and flexible.

[0087] 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 method for predicting the status of IoT devices and elastically scheduling resources in a digital space, characterized in that, The method specifically includes: Collect operational monitoring data of IoT devices and current remaining resource values ​​within the digital space; Based on the operation monitoring data of IoT devices, the status of IoT devices is predicted, and the predicted value of the IoT device status stress is determined. The basic priority score of IoT devices is determined based on the predicted stress level of IoT devices and the current remaining resource value. The physical connection topology, data workflow dependencies, and historical fault concurrency statistics of IoT devices within the digital intelligence space are collected. Based on these data, an IoT device relationship network is constructed. This network is a directed acyclic graph, with nodes connected by directed edges. Obtain the location of each node in the IoT device relationship network, and determine the IoT device network hierarchy depth based on the location of each node in the IoT device relationship network; Determine the priority reduction coefficient of IoT devices based on the network layer depth of IoT devices; The final priority score of the IoT device is determined based on the IoT device priority reduction coefficient and the IoT device basic priority score. Based on the final priority score of IoT devices and the predicted stress level of IoT devices, resources for IoT devices are scheduled.

2. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 1, characterized in that, The determination of the predicted stress level of IoT devices specifically includes: Through the formula: ; Determine the predicted value of the status tension of IoT devices ; In the above formula, This refers to the predicted stress level of the i-th IoT device at time t+1 in the future. This refers to the standardized value of the IoT device operation monitoring data at time t-k+1 for the i-th IoT device. This refers to the weight coefficient corresponding to the k-th historical moment, where k refers to the number of historical moments.

3. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 2, characterized in that, The determination of the basic priority score for IoT devices: Through the formula: ; Determine the basic priority score for IoT devices ; In the above formula, This refers to the basic priority score of the i-th IoT device. This refers to the predicted stress level of the i-th IoT device at time t+1 in the future. This refers to the current remaining resource value of the i-th IoT device. This refers to the maximum resource value of the i-th IoT device. All are weighting coefficients, and .

4. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 1, characterized in that, Determining the network layer depth of IoT devices specifically includes: Through the formula: ; Determine the network layer depth of IoT devices ; In the above formula, This refers to the network layer depth of the i-th IoT device. This refers to the network layer depth of the j-th IoT device, and E refers to the set of directed edges from node j to node i. In this context, the depth of the central source node in the IoT device relationship network is defined as 0, and the central source node refers to the root node with an in-degree of 0.

5. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 4, characterized in that, The determination of the priority reduction coefficient for IoT devices specifically includes: Through the formula: ; Determine the priority reduction factor for IoT devices ; In the above formula, This refers to the network layer depth of the i-th IoT device. This refers to the maximum network hierarchy depth of all nodes in the interconnected network of IoT devices. This refers to the adjustment factor that affects the outcome. This is a constant term.

6. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 1, characterized in that, The determination of the final priority score for IoT devices specifically includes: Through the formula: ; Determine the final priority score for IoT devices ; In the above formula, This refers to the basic priority score of the i-th IoT device. This refers to the priority reduction coefficient of the i-th IoT device.

7. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 6, characterized in that, The specific conditions for determining the final priority score of IoT devices include: ; in, This refers to the arithmetic mean of the basic priority scores of all IoT devices. This refers to the deviation threshold; If the conditions are not met when determining the final priority score of IoT devices, the priority reduction coefficient for IoT devices will be adjusted. The value is set to 0.

8. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 7, characterized in that, The deviation threshold The specific methods of obtaining it include: Through the formula: ; Determine the available ratio of resources for scheduling ; In the above formula, This refers to the characteristic value of currently schedulable resources. This refers to the maximum value of the scheduled resources; Based on the mapping function, the resource scheduling availability ratio is calculated. Mapped to deviation threshold .

9. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 1, characterized in that, The process of scheduling IoT device resources based on the final priority score and predicted state stress of IoT devices specifically includes: Based on the final priority score of IoT devices, the IoT devices are ranked to determine the priority sequence for IoT device resource scheduling; Based on the priority sequence of IoT device resource scheduling, resource scheduling is performed on IoT devices according to the predicted value of IoT device status tension.

10. The method for predicting the state of IoT devices and elastically scheduling resources for digital intelligent spaces according to claim 9, characterized in that, The method of scheduling resources for IoT devices based on the priority sequence of IoT device resource scheduling and according to the predicted value of IoT device status tension includes: Based on the predicted stress levels of IoT devices, a set of IoT devices that require resource scheduling is selected. Based on the priority sequence of IoT device resource scheduling, resource scheduling is performed on IoT devices in the set of IoT devices that need resource scheduling.

Images